CN101076004A - Wireless communication device - Google Patents

Abstract

The invention provides a wireless communication device for receiving a package constituted by signal modulated by OFDM. A band-pass filter pick up OFDM signal of desired frequency band. The desired OFDM signal is amplified by low-pass amplifier based on the gain controlled by the received signal intensity. A frequency converter down-converses the FDM signal into baseband signal. An A/D converter converses the baseband signal into digital signal. A first high pass filter removes DC deviation from baseband signal corresponding to preset preamble of the package. A frequency shift etimator estimates frequency shift from the sampled signal which has removed DC deviation from baseband signal by the first high pass filter. A frequency shift corrector removes the estimated frequency shift from the baseband signal. A demodulator demodulates the subcarrier signal in the frequency domain from baseband signal which is compensated with frequency shift.

Description

Radio communication device

The cross reference of related application

The theme of JP2007-037719 that the present invention is contained in the Japanese patent application JP 2006-137047 that submitted to Japan Patent office on May 16th, 2006, submitted to Japan Patent office on February 19th, 2007 and the JP2007-108046 that submits to Japan Patent office on April 17th, 2007, their full content is hereby expressly incorporated by reference.

Technical field

The present invention relates to be used for receiving the radio communication device of radio frequency (RF) signal of modulating by OFDM (OFDM).Especially, the present invention relates to be used for using do not utilize the direct converting structure of intermediate frequency (IF) level to come the radio communication device of received signal.

More specifically, the present invention relates to a kind of radio communication device that is used for using the training sequence that adds packet header to remove frequency shift (FS) and demodulating ofdm symbol.Especially, the present invention relates to a kind of being used for exists time dependent DC skew and homophase and the quadrature accurate radio communication device of estimated frequency skew under (IQ) unbalanced situation mutually at the OFDM symbol that is received.

Background technology

Wireless network causes concern as the no cable system that replaces traditional wired communication system.IEEE (Institute of Electrical and Electric Engineers) the 802.11st, the standard that wireless network is general.

For example, when in the environment out of doors wireless network being set, the problem that receiving system receives direct wave and a plurality of reflected wave and postpones the stack of ripple can occur, that is, multipath reception occur.Multipath reception causes delay distortion (or frequency selective fading), thereby causes communication error.Delay distortion causes intersymbol interference.In WLAN (wireless local area network) (LAN) standard such as IEEE 802.11a/g, employing as a kind of OFDM modulating mode of multi-carrier modulation pattern (for example, referring to IEEE 802.11a, part 11: WLAN media interviews controls (MAC) layer and physical layer (PHY) illustrate: the high-speed physical layer in the 5GHZ frequency band; And IEEE 802.11g, part 11: WLAN media interviews controls (MAC) layer and layer physics (PHY) illustrate: the high-speed physical layer in the 2.4GHZ frequency band).

Form

Standard	IEEE 802.11a	IEEE 802.11g
			Frequency of utilization	5.2GHz	2.4GHZ
Maximum transfer speed	54Mbps	54Mbps
			Modulating mode	OFDM	OFDM

The OFDM transmitter will convert parallel data to the speed that is lower than the rate of information throughput by the serial signal information transmitted at each symbol period, then a plurality of parallel data streams be distributed to subcarrier, be used for the amplitude and the phase place of each subcarrier are modulated.The OFDM transmitter is also carried out reverse inverse fast Fourier transform (IFFT) to a plurality of subcarriers, converting frequency domain subcarrier to time-domain signal, and transmits resulting signal.The OFDM receiver is carried out the operation opposite with the operation of OFDM transmitter.That is, the OFDM receiver is carried out fast Fourier transform (FFT) to convert time-domain signal to frequency-region signal, is used for carrying out demodulation according to the modulating mode corresponding to subcarrier.Conversion is also carried out also-gone here and there to the OFDM receiver, and regeneration is by the raw information of serial signal transmission.Determine the frequency of carrier wave so that subcarrier is orthogonal on symbol period.Orthogonal subcarrier means that the peak point of the frequency spectrum of given subcarrier is matched with the zero point of the frequency spectrum of other subcarrier consistently, and can not occur between them crosstalking.Therefore, the transmission data are to transmit having on a plurality of carrier waves of orthogonal frequency, and have realized narrow bandwidth, high-frequency service efficiency and the advantage high to the resistance of frequency selective fading of carrier wave.Therefore, can realize effective OFDM modulator-demodulator by using fft algorithm.The OFDM transmission mode is used in Wireless LAN system, such as the ground digit broadcasting system (for example, referring to J.OLSSON, " WLAN/WCDMA Dual-Mode Receiver Architecture DesignTrade-Offs " Proc.of IEEE 6th CAS Symp., vol.2, pp.725-728, on May 31st, 2004 was to June 2), the 4th third-generation mobile communication system, and in various other wideband digital communication systems of power-line carrier communication system.

In the RF of radio communication device front end, in transmission course, usually after frequency of utilization transducer (quadrature modulates device) up-converts to analog baseband signal the RF band signal and uses the band pass filter restricted band, use gain-changeable amplifier circuit to amplify through-put power.In receiving course, amplify the signal that receives by antenna by low noise amplifier (LNA), use local frequency f then _LCIt is down-converted to baseband signal.Automatic gain control (AGC) circuit is used to make the electric current of self signal to maintain suitable constant level.

In recent radio communication device, the frequency converter that transmitting/receiving signal is carried out up-conversion or down-conversion uses direct converting structure, to use carrier frequency f _CAs local frequency f _L0Carry out direct frequency inverted.Directly converting structure does not use outside intermediate frequency (IF) filter (being also referred to as " RF inter-stage filter "), has reduced size and power consumption and has increased integration thereby compare with super-heterodyne architecture.In addition, in principle, do not generate alias, and directly converting structure is outstanding at the design aspect of transmitter and receiver.Yet, directly changing in the receiver architecture, so pointed out because receive frequency equates with local frequency by the self-mixing of local signal to cause that in output place of low-converter problem that DC component or DC be offset (for example, referring to Anuj Batra, " 03267r1P802-15_TG3a-Multi-band-OFDM-CFP-Presentation.ppt ", pp.17, in July, 2003).Owing to self-mixing appears in the limited isolation between the RF port of local signal and low noise amplifier or frequency mixer.Term DC used herein is defined as the 0Hz (zero IF) in the baseband signal in the OFDM modulating mode.

Ofdm communication system has the problem that the little error between the oscillator frequency in the transmitter and receiver (for example, in WLAN, using the oscillator with about 20ppm precision) can cause the skew of receiver medium frequency.Though subcarrier does not disturb each other, there is the frequency orthogonal that can't keep under the situation of frequency shift (FS) between the subcarrier, thereby causing the deterioration of demodulation characteristics, that is, receive the error in the data.

In packet switched wirelss communication system such as IEEE 802.11 communication systems, the symbol known to place, the packet header placement transmitter and receiver of each bag, that is, and training sequence.Receiver uses the training sequence that receives to carry out the automatic gain control of low noise amplifier, DC skew estimation and removal, Frequency offset estimation and removal, bag and detects and regularly detect.Handle if carry out frequency shift (FS), will increase the complexity and the power consumption of circuit structure so by analog circuit.Therefore, the present inventor considers estimation and the compensation deals of preferably carrying out frequency shift (FS) by digital processing.In response to the observation of frequency shift (FS), data phase is inverted with correcting frequency shift.

Below will under the background of IEEE 802.11a/g, check the problem of frequency shift (FS).Figure 15 show preamble structure therefor specified in IEEE 802.11a/g (for example, referring to M.Itami, " OFDM Modulation Technique ", Triceps 2000; And IEEE802.11a, part 11: WLAN media interviews controls (MAC) and physical layer (PHY) illustrate: the high-speed physical layer in the 5GHZ frequency band).As shown in figure 15, the short preamble period of 8.0 μ s and the long preamble period of 8.0 μ s are added in the packet header.Short preamble period is made of short training sequence (STS), wherein, repeats to transmit ten short preamble symbol t ₁To t ₁₀Long preamble period is made of long training sequence (LTS), wherein, repeats to transmit two long preamble symbol T after the GI2 of the protection interval of 1.6 μ s ₁To T ₂A short preamble symbol is made of 12 subcarriers, and has the length of 0.8 μ s, and this is corresponding to the IFFT/FFT period T _FFT1/4th.A long preamble symbol is made of 52 subcarriers, and has the length of 3.2 μ s, and this is corresponding to the IFFT/FFT period T _FFTAs shown in figure 30, ofdm signal does not comprise DC or 0Hz subcarrier, to avoid the DC offset interference.

IEEE 802.11a/g does not specify the use of preamble.Usually, receiver is provided with the gain of receiver and uses four STS symbol correction DC skews of 0.8 μ s, and the estimation and correction, the bag that use remaining six STS symbols to carry out frequency shift (FS) detect and slightly regularly detect.

According to following equation (1), can use the STS in 0.8 microsecond cycle to obtain the estimation of frequency shift (FS):

$\mathrm{\Δf}\left(k\right)=\frac{1}{2\mathrm{\π}{T}_{\mathrm{STS}}}\frac{1}{M}\underset{i=k}{\overset{k+M-1}{\mathrm{\Σ}}}\arg \left(S\left(i\right)S{(i-16)}^{*}\right)\·\·\·\left(1\right)$

Wherein, T _STSRepresent 0.8 microsecond, S (i) is illustrated in the STS signal of 20MHz frequency place sampling, S ^*(i) complex conjugate of expression STS signal, and M represents the sample mean number.

Can detect and slightly regularly detect based on carrying out bag by the standardized corrected value of STS power level, this be provided by following equation (2):

$\mathrm{CF}\left(k\right)=\frac{1}{N}\underset{i=k}{\overset{k+M-1}{\mathrm{\Σ}}}\frac{S\left(i\right)S{(i-16)}^{*}}{S\left(i\right)S{\left(i\right)}^{*}}\·\·\·\left(2\right)$

Wherein, N represents the sample mean number.

In the process that bag detects, the corrected value that the above equation (2) of serving as reasons provides is provided with threshold level, and detects bag when corrected value surpasses threshold level.In the thick process that regularly detects, utilize corrected value to become the characteristic that reduces from increase at the end of STS.That is, the corrected value of current detection and the corrected value of before having determined are compared, to determine thick timing.

Therefore, receiver uses automatic gain control, DC skew estimation and removal, Frequency offset estimation and removal, the bag of the preamble portion execution low noise amplifier of each bag to detect and regularly detection.

Yet the precision that Frequency offset estimation, bag detect and regularly detect is very sensitive to the DC skew, and appears at and have the problem that is difficult to accurate estimated frequency skew under the DC drift condition.Especially in above-mentioned direct converting structure, the problem that the DC that is caused by self-mixing is offset is very serious, and the quality of received signal can be weakened by frequency shift (FS) and DC skew.

For example, there is under the situation of DC skew corrected value even can during dead time, increase in I axle and Q axle input.That is, corrected value is constant to add one, and owing to increase continuously, corrected value surpasses the threshold value that detects the basis as bag.Therefore, receiver identifies even also receives bag in the dead time section, thereby causes operate miss.

In addition, exist under the situation of DC skew, even in the incoherent each other part of received signal, reduce because the influence of DC skew makes corrected value not become from increase at I axle and Q axle input.Therefore, deterioration thick timing detect characteristic.

In addition, under the situation that has the DC skew, the precision of Frequency offset estimation reduces, and the characteristic of the further deterioration received signal of residual frequency shift (FS).The frequency shift (FS) of Qu Chuing does not cause the phase place rotation of all subcarriers of training sequence OFDM symbol afterwards yet, even and cause that noise (SN) the error flat bed of error still can occur wrapping than increase.On the contrary, when estimating the DC skew under the situation that has frequency departure, be difficult to accurately estimate the DC skew.Therefore, expectation solves the problem that has DC skew and frequency shift (FS).

As mentioned above, be desirably in and preceding four STS symbol t ₁To t ₄Equally in Duan time period, in the previous stage of frequency offset correction circuit module, carry out high-precision DC offset correction.Usually, be difficult to realize short-time high-accuracy DC offset correction, and power consumption and circuit size increase significantly.

Existed be used to use high pass filter (HPF) remove the method for DC skew, be used for estimating simultaneously DC skew and frequency shift (FS) method, be used for the parallel DC of estimation skew and frequency shift (FS) method, be used for that repetition DC skew is estimated and method of frequency offset compensation or the like.

Figure 16 diagram show use HPF remove the DC skew receiver structure (for example, referring to W.Namgoong and T.H.Meng, " Direct-Conversion RFReceiver Design) ", IEEE Trans.on Commun.Vol.49, No.3, March calendar year 2001).In receiver shown in Figure 16, use HPF to remove to be included in the DC offset component in the OFDM symbol of receiving.Then, carry out signal processing and be offset, and remove frequency shift (FS) the OFDM symbol after training sequence with estimated frequency.Yet in the method, HPF makes the nearly DC signal attenuation in the OFDM symbol, possible deterioration demodulation characteristics.

A kind of method of nearly DC signal attenuation that prevents is for fully reducing the cut-off frequency f of HPF with respect to subcarrier interval _c(referring to Figure 17 A).Yet, if gain by automatic gain control break low noise amplifier, the problem of change DC skew (for example when then having generation, referring to S.Otaka, T.Yamaji, R.Fujimoto and H.Tanimoto, " A Low Offset1.9GHz Direct Conversion Receiver IC with Spurious Free DynamicRange of over 67 dB ", IECE Trans.on Fundamentals, vol.E84-A, no.2, pp.513-519, February calendar year 2001).Has low cut-off frequency f _cHPF have low-response, and become the DC skew can transmit by HPF the time.

For example, in preamble structure therefor shown in Figure 15, the gain of low noise amplifier changes around the center of short preamble period from high to low.The DC skew greatly changes in time according to the change that gains, and high fdrequency component is included in the DC skew (referring to the part (a) of Figure 31).Owing to have low cut-off frequency f _cHPF have low-response, so the time become the DC skew high fdrequency component transmit by HPF, and in the level of back through frequency offset estimator.If this residual DC still is present in the follow-up long preamble period, then can influences the thin Frequency offset estimation of in long preamble period, carrying out (part (b) of ginseng Figure 31), thereby cause lower estimated accuracy.

For example, in IEEE 802.11a/g, preamble period obviously shortens, and expectation uses HPF to restrain the residual DC skew fast.By greatly increasing the cut-off frequency f of HPF _c(for example minimize convergence time, referring to T.Yuba and Y.Sanada, " DecisionDirected Scheme for IQ Imbalance Compensation on OFDCM DirectConversion Receiver ", IEICE Trans.on Communications, vol.E89-B, no.1, pp.184-190, in January, 2006).

Cut-off frequency f with increase _cHPF the variation of the DC skew that causes by the gain that changes low noise amplifier is had high response, even can be by effectively nearly DC signal (referring to Figure 17 B).Therefore, possible deterioration OFDM demodulation characteristics.

Consider better transient response and fast convergence rate, preferably the cut-off frequency f of HPF _cVery high.In preamble structure therefor shown in Figure 15, at the DC component with have between the subcarrier near the DC component among the 1.25MHz STS at interval, can be even cut-off frequency is very high by the signal of nearly DC subcarrier yet.Yet, having between the subcarrier near DC among the 312.5kHz follow-up LTS at interval, HPF will be by the signal of nearly DC subcarrier, thereby causes the deterioration of demodulation characteristics.

The structure that Figure 18 diagram shows the receiver of estimating DC skew and frequency shift (FS) simultaneously (for example, referring to G.T.Gil, I.H.Sohn, J.K.Park and Y.H.Lee, " JointML Estimation of Carrier Frequency; Channel, I/Q Mismatch, and DCOffset in Communication Receivers ", IEEE Trans.on Vehi.Tech., Vol.54, No.1, in January, 2005).In receiver shown in Figure 180, use the maximum likelihood estimation technique to estimate simultaneously and compensate DC to be offset and frequency shift (FS).Yet, owing to a large amount of calculating of the maximum likelihood estimation technique and long computing time make that being difficult to carry out maximum likelihood in the system with limited migration time estimates.In preamble structure therefor shown in Figure 15, be desirably in preceding approximately four STS symbol t ₁To t ₄, that is, finish the DC skew in about 3.2 microseconds and estimate.

Figure 19 diagram show the parallel DC of estimation skew and frequency shift (FS) receiver structure (for example, referring to C.K.Ho, S.Sun and P.He, " Low Complexity FrequencyOffset Estimation in the Presence of DC Offset ", Proc.of IEEEInternational Conference on Communications 2003, Vol.3, pp.2051-2055, in May, 2003; And U.S. Patent Application Publication the 2003/0174790th and No. 2005/0078509).The DC offset estimator is on average estimated the DC skew by asking on whole preamble.Frequency offset estimator in the level of back calculates the correlation function of preamble, and deducts the DC skew of estimation, estimates accurate frequency shift (FS) the signal to remove from the DC skew.Yet if at the level of preamble reception period by the change DC skews such as gain of change low noise amplifier, the DC skew is estimated in possible errors.

Figure 20 diagram shows that repetition DC skew is estimated and the structure of the receiver of frequency offset compensation No. the 2005/0020226th, 2003/0133518,2004/0202102 and 2005/0276358, U.S. Patent Publication (for example, referring to).In receiver shown in Figure 20, DC skew remover is removed the DC skew, estimated frequency skew then.After the compensating frequency skew, estimate that further the DC skew is to leave out residual DC skew.This method is offset the feedback loop of removing with restraining DC for a long time, and is difficult to be used for the method for short preamble.In addition, if the change DC skews such as gain by changing low noise amplifier then error can occur in Frequency offset estimation.

As everyone knows, the DC offset level changes according to the change of the gain that is provided with in low noise amplifier.Figure 18 does not fully take into account the DC skew to receiver shown in Figure 20 to be changed and changes in time according to the gain in the low noise amplifier.

OFDM directly changes the problem that receiver also has the IQ imbalance and is offset by the caused DC of the self-mixing of local signal.Directly converting structure does not use the IF signal in numeric field, and does not carry out the IQ quadrature modulates in numeric field and in analog domain.The IQ imbalance is uneven caused by between homophase (I) component and quadrature phase (Q) component.Particularly, IQ imbalance mutually is caused by the non-phase quadrature between the local signal that inputs to I channel and Q channel mixer, and IQ gain imbalance is caused (for example by the gain inequality between the signal in I channel and the Q channel, referring to T.Yuba and Y.Sanada, " Decision Directed Scheme for IQ Imbalance Compensation onOFDCM Direct Conversion Receiver ", IEICE Trans.onCommunications, vol.E89-B, no.1, pp.184-190, in January, 2006).Similar with the DC skew, the IQ imbalance causes the deterioration of Frequency offset estimation precision, and influences decoding characteristics.

Therefore, directly change in the receiver architecture at OFDM, the quality of received signal can reduce owing to frequency shift (FS), DC skew and IQ imbalance.When not considering, the receiver shown in Figure 18 to Figure 20 becomes DC skew and IQ imbalance.

Summary of the invention

Therefore, expectation provides a kind of fabulous radio communication device that uses direct converting structure, wherein, can become when existing and suitably remove frequency shift (FS) under the situation of DC skew and modulate to realize the OFDM with higher characteristic.

In addition, also expectation provides a kind of fabulous radio communication device, wherein, becomes when existing simultaneously in the OFDM symbol that receives under the situation of DC skew, IQ imbalance and frequency shift (FS), can remove DC skew and accurately estimated frequency skew.

According to the first embodiment of the present invention, provide a kind of radio communication device that is used for receiving the bag that constitutes by signal by OFDM (OFDM) modulation.This radio communication device comprises following element.Band pass filter extracts the ofdm signal of desired frequency band.Low noise amplifier with gain of controlling according to the intensity of received signal amplifies the ofdm signal of desired frequency band.Frequency converter down-converts to baseband signal with the FDM signal that amplifies.Analog to digital converter converts baseband signal to digital signal.First high pass filter is removed the DC skew from the predetermined preamble baseband signal partly corresponding to bag.Frequency offset estimator estimated frequency skew from the sampled signal of forming baseband signal (from this baseband signal, removing the DC skew by first high pass filter).Frequency offset corrector is removed estimated frequency shift (FS) from baseband signal.Demodulator demodulation from the baseband signal of compensating frequency skew is arranged in the subcarrier signal in the frequency domain.

Embodiments of the invention relate to the radio communication device that is used to use direct converting structure reception ofdm signal.Do not have the direct converting structure of IF filter to realize broadband receiver easily, and allow the flexibility of receiver design.Yet, have DC bias effect frequency shift (FS) that the self-mixing by local signal causes or the problem that regularly detects.

In radio communication device, using first high pass filter from corresponding to the predetermined preamble baseband signal part partly of bag, after the removal DC skew, to be offset with the high accuracy estimated frequency according to the embodiment of the invention.Remove estimated frequency shift the receiving baseband signal part after the part of estimated frequency skew.

Difference filter with simple circuit structure is used as first high pass filter and removes the DC skew.This difference filter has very high response to the change of the DC offset level that causes in the gain of preamble reception period by changing low noise amplifier etc.

If be used for the cut-off frequency f that the DC skew is removed _cIncrease, then the response to the change of the DC skew that causes by the gain that changes low noise amplifier increases.Yet, exist even can cause the deterioration (referring to Figure 17 B) of demodulation characteristics by the problem of nearly DC signal.Yet, be divided and carry out DC skew removal owing to be used for the preamble portion of Frequency offset estimation, so the cut-off frequency f that increases _cMay not influence the demodulation characteristics in the follow-up data part.

If DC skew is because the change of the gain of low noise amplifier etc. and changing fast, then the DC skew may be transmitted by difference filter.Therefore, in case can be configured to detect the quick change of DC skew as the difference filter of first high pass filter time, just detection signal is inputed to the frequency offset estimator in the back grade.When the input detection signal, frequency offset estimator is not carried out Frequency offset estimation to the sampled signal that obtains from difference filter.Thereby, realized high estimated accuracy.

Suppose that the ofdm signal that inputs to radio frequency communication devices does not comprise the DC subcarrier.Frequency offset estimator can be configured to use the preamble estimated frequency skew of two OFDM symbols of transmission.

Particularly, an OFDM symbol is made of n subcarrier.When i sampling of the time waveform of the OFDM of two transmission symbol represented by s (i), the sampling of the OFDM symbol of first transmission is by { s (0), s (1), ..., s (n-1) } expression, the sampling of the OFDM symbol of second transmission is by { s (n), s (n+1), ..., s (2n-1) } expression, frequency shift (FS) represented by Δ f, and the DC skew is when being represented by D, first high pass filter can be configured to the baseband signal that receives that is provided by following equation (3) is carried out the operation that is provided by following equation (4), and output sampled signal d (i):

r(i)＝s(i)exp(j2πΔfi)+D …(3)

d(i)＝r(i+1)-r(i)

＝s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi) …(4)

Frequency offset corrector can be configured to use sampled signal d (i) to carry out by the given operation of following equation (5), to estimate to be included in the frequency shift (FS) Δ f among the baseband signal r (i) that is received:

$d(i+n)/d\left(i\right)=\frac{s(i+1+n)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+1+n)\right)-s(i+n)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+n)\right)}{s(i+1)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+1)\right)-s\left(i\right)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}\left(i\right)\right)}$

$=\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\left(n\right)\right)\&CenterDot;\&CenterDot;\&CenterDot;\left(5\right)$

In as IEEE802.11a/g, have the short preamble that the transmission of same-sign repeats and grow the Bao Touchu that preamble all is included in each bag according to an example of the wireless communication system of the embodiment of the invention.Therefore, by carrying out aforesaid operations, can estimated frequency skew more accurately from the signal of removing the DC skew.

The sampled signal d (i) that exports from first high pass filter that is made of difference filter is determined by above-mentioned equation (4).More specifically, given as following equation (6), will be added into sampled signal d (i) by the DC offset variation between i that D (i+1)-D (i) provides and (i+1) individual sampling.If DC offset variation amount is very little, then can not go wrong.Yet, if DC skew is because gain change of low noise amplifier etc. and greatly change DC bias effect Frequency offset estimation between sampled signal:

d(i)＝r(i+1)-r(i)

＝{s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi)}+{D(i+1)-D(i)}

…(6)

Therefore, when the absolute value of sampling output d (i) because by D (i+1)-D (i) provides between i and (i+1) individual sampling DC offset variation and when big, first high pass filter can be configured to export detection signal in back grade frequency offset estimator.In response to detection signal, frequency offset estimator does not use the operation that is provided by above-mentioned equation (5) that i sampling output is carried out frequency shift (FS) according to estimates.Thereby, realized high estimated accuracy.

As mentioned above, frequency offset estimator estimated frequency skew from the signal of removing the DC skew.Remove by the frequency offset estimator estimated frequency shift in the receiving baseband signal part (that is, therefrom not removing the DC skew) of frequency offset corrector after the part that is used for Frequency offset estimation.In other words, only the preamble portion of carrying out Frequency offset estimation is removed the DC skew, and the preamble portion after the part that is used for Frequency offset estimation and payload portions are removed the skew of frequency rate but do not remove the DC skew.In this case, carry out Frequency offset estimation even may produce with high accuracy, the DC skew still can influence the problem of the demodulator circuit in following stages.

Therefore, preferably, also at the device that in institute's receiving baseband signal, is provided for removing the DC skew in the part after the part of estimated frequency skew.

For example, use at frequency converter under the situation of direct converting structure, the phase place of the local frequency by local oscillator vibration can be according to being reversed by the frequency offset estimator estimated frequency shift.Therefore, can from the receiving baseband signal part that after the part of estimated frequency skew, comprises the DC skew, eliminate the influence of DC skew and frequency shift (FS).

Alternatively, radio communication device can also comprise the DC offset estimator, estimates the skew by the DC in the digital baseband signal of analog to digital converter conversion, and can remove the DC skew of estimating from the data converted baseband signal.

Usually by being asked, the digital baseband signal of conversion on average estimates the DC skew.Therefore, if DC skew because gain change of low noise amplifier etc. and changing fast, then mean value is die on, and the formation estimated of DC skew will cause the deterioration of precision.Therefore, in case detect the quick change of DC skew, difference filter can be configured to detection signal is inputed to the DC offset estimator.The DC offset estimator can be configured to get rid of the data estimator of estimating and reappraise the DC skew before the input detection signal, to prevent the reduction of estimated accuracy.

Alternatively, can carry out filtering to the receiving baseband signal that uses direct converting structure down-conversion, and can remove caused DC skews such as self-mixing, after this, the conversion of signals that produces can be become digital signal by local signal by second high pass filter.Second high pass filter allows all signals to pass through, and preferably the cut-off frequency of second high pass filter is provided with relatively low, make the nearly DC signal in the OFDM symbol.

Second high pass filter with low cut-off frequency has low-response, and causes DC skew influence for a long time.Yet first high pass filter that has enough higher cutoff frequencies in use is removed DC skew estimated frequency skew afterwards, thereby obtains high estimated accuracy.In addition, only the preamble portion of carrying out Frequency offset estimation is carried out the filtering of higher cutoff frequency, and demodulation characteristics that can the deterioration follow-up signal.

Usually, not only in Frequency offset estimation but also in other signal processing that detects or slightly regularly detect such as bag, deterioration in characteristics is very responsive to the DC skew.

Therefore, radio communication device can also comprise that signal that use has therefrom been removed DC skew by difference filter carries out that bag detects and the thick detector that regularly detects.

Radio communication device can also comprise the switch that the output of analog to digital converter is connected to specially the path of the path of guiding first high pass filter or the DC offset corrector that leads.Switch can be configured to the output of analog to digital converter is transformed into the path of guiding DC offset corrector from the path of first high pass filter that leads when detector detects the predetermined preamble part that is used for Frequency offset estimation terminal.

Owing in the cycle before detecting predetermined preamble part end the output of analog to digital converter is connected to the path of guiding first high pass filter, frequency offset estimator can use the receiving baseband signal that has used first high pass filter therefrom to remove the DC skew to be offset with the high accuracy estimated frequency in a period of time before the end of predetermined preamble part.

The DC offset estimator can be configured to estimate in a period of time before predetermined preamble part end the DC skew.When detecting predetermined preamble part terminal, the output of analog to digital converter switches to the path of guiding DC offset corrector.Can use the DC skew of in long-time section, estimating by the DC offset estimator, the terminal receiving baseband signal afterwards of predetermined preamble part is partly carried out the DC offset correction with high accuracy.In the receiving baseband signal part after the predetermined preamble part is terminal, do not use first high pass filter to remove the DC skew, and do not consider the deterioration of SNR feature.

Frequency offset estimator can be configured to estimated frequency skew in a period of time before predetermined preamble part end, and frequency offset estimator can be configured to proofread and correct estimated frequency shift the receiving baseband signal part after the predetermined preamble part is terminal.

According to embodiments of the invention, provide a kind of wireless communication system of the IEEE of meeting 802.11a standard.In IEEE 802.11a standard, add in the packet header of each bag with the short preamble portion that constitutes by short training sequence and by the long preamble portion that the long training sequence with less relatively subcarrier interval constitutes with relatively large subcarrier interval.

Configuration is according to the feasible short preamble portion with relatively large subcarrier interval that utilizes of the radio communication device of the embodiment of the invention, and the difference filter that has enough higher cutoff frequencies in use is removed DC skew estimated frequency skew afterwards, thereby obtains high estimated accuracy.Promptly, owing to only use short preamble portion to carry out frequency offset correction, go out the path that the output of analog to digital converter is switched to guiding DC offset corrector from the path of first high pass filter that leads so switch can be configured to the beginning of the long preamble portion after short preamble portion end.

Then, frequency offset estimator is the estimated frequency skew in short preamble portion, and frequency offset corrector is removed estimated frequency shift from long preamble portion.The DC offset estimator is estimated the DC skew in short preamble portion, and the DC offset corrector is removed the DC skew of estimating from long preamble portion.

The long preamble portion of transmission can be received the machine use with arbitrary form after short preamble portion.Usually, use short preamble portion to carry out the coarse frequency skew and estimate, use long preamble portion to carry out thin frequency offset correction and channel estimating then.Therefore, in short preamble portion, estimate accurate frequency shift (FS), to realize that precise channels is estimated more.

For example, radio communication device can also comprise: second frequency offset estimator, estimated frequency skew in the long preamble portion after short preamble portion; And the second frequency offset corrector, in long preamble portion, remove by second frequency offset estimator estimated frequency shift.The second frequency offset estimator is received in the receiving baseband signal part after the long preamble portion of therefrom removing estimated frequency shift and DC skew in short preamble portion, and estimated frequency is offset.Remove by second frequency offset estimator estimated frequency shift the receiving baseband signal part of second frequency offset corrector after long preamble portion

Alternatively, receiving baseband signal after the long preamble portion of therefrom removing estimated frequency shift and DC skew in short preamble portion partly can be fed back to frequency offset estimator, with from long preamble portion receiving baseband signal afterwards partly estimated frequency be offset.Remove estimated frequency shift the receiving baseband signal part of frequency offset corrector after long preamble portion.

In either event, estimated frequency shift and DC skew and proofreaied and correct in the receiving baseband signal of the residual frequency offset will of estimating in the part after long preamble portion and estimated channel from remove down short preamble portion, thus acquisition has high-precision channel information.

If the baseband signal that is received comprises the IQ imbalance, even then carry out frequency offset correction, the receiving feature that can not obtain to expect.For fear of this unfavorable, radio communication device can also comprise uneven estimator of IQ and IQ disequilibrium regulating device.By this structure, can delete the IQ imbalance that is included in institute's receiving baseband signal, and can obtain the further receiving feature of improvement.

OFDM directly changes receiver not only to have DC skew but also has by phase difference between the local signal that inputs to I axle and Q axle frequency mixer and the unbalanced problem of the caused IQ of the difference of vibration between the frequency mixer.Similar with the DC skew, the IQ imbalance causes the deterioration of Frequency offset estimation precision, and influences decoding characteristics.

When the DC skew was therefrom removed in the frequency offset estimator use and wherein still exist the unbalanced receiving baseband signal estimated frequency of IQ to be offset, frequency offset information comprised frequency offseting value Δ f and the component that is brought by the IQ imbalance, that is, and and the IQ unbalanced component.

In general communication system, transmit a plurality of preamble symbol from transmitter, and the frequency offset estimator in the receiver can be estimated the frequency shift (FS) of each preamble symbol.Frequency shift (FS) can be expressed as the vector on the complex space.The vector direction of expression IQ unbalanced component is according to preamble symbol and difference.By with the addition of preamble symbol estimated frequency shift order, frequency offset estimator can reduce to be included in the IQ unbalanced component in the estimated frequency shift value relatively, and can finally obtain accurate more frequency shift (FS).

Receiver generally includes the gain controller of regulating the low noise amplifier gain.When during the Frequency offset estimation of carrying out by frequency offset estimator, changing the gain of low noise amplifier, be included in the IQ unbalanced component increase in estimated frequency shift when big the gain is set.Therefore, if will then may be difficult to reduce effectively the ratio of IQ unbalanced component to the simple addition of a plurality of preamble symbol estimated frequency shift.

In this case, when receiving corresponding preamble symbol, frequency offset estimator can will be weighted each frequency shift (FS) that preamble symbol is estimated according to the gain that is arranged in the low noise amplifier, and the frequency shift (FS) of weighting can be obtained final frequency offseting value mutually.Therefore, the IQ unbalanced component in the estimated frequency shift value can be relatively reduced to be included in, and accurate more frequency shift (FS) can be finally obtained.

Usually, in receiver, when input begins, determine the big gain of low noise amplifier, and lower gain is changed in gain according to the power of received signal.Therefore, frequency offset estimator is applied to estimated frequency shift in first some preamble symbol (during big gain is set) with little weight in low noise amplifier, and big weight is applied to change into estimated frequency shift in the subsequent preamble symbol of less gain in gain, and the frequency shift (FS) of weighting is obtained final frequency shift (FS) mutually.Particularly, weight being applied to frequency shift (FS) is equivalent to the Frequency offset estimation vector (following description) or the output of difference filter be multiply by weighted factor.

Particularly, frequency offset estimator is configured to calculate weighted factor based on the absolute value to each preamble symbol estimated frequency shift, frequency shift (FS) being weighted, and the frequency shift (FS) of weighting obtained final frequency offseting value mutually by weighted factor.

For example, frequency offset estimator can surpass the frequency shift (FS) of predetermined threshold and weighted factor 1 is applied to the frequency shift (FS) that absolute value is no more than predetermined threshold and will be weighted the preamble symbol estimated frequency shift by weighted factor 0 being applied to absolute value, and the frequency shift (FS) of weighting can be obtained final frequency offseting value mutually.That is, ignore estimated frequency shift in the preamble period that big gain is set in low noise amplifier.For example, can determine predetermined threshold based on the intensity of received signal.

Alternatively, frequency offset estimator can will be weighted the preamble symbol estimated frequency shift by being applied to frequency shift (FS) by the reciprocal formed weighted factor of the absolute value of frequency shift (FS), and the frequency shift (FS) of weighting can be obtained final frequency offseting value mutually.

According to embodiments of the invention, can realize using the fabulous radio communication device of direct converting structure, wherein, can suitably remove the OFDM demodulation that frequency shift (FS) has higher characteristic with realization under the situation that becomes the DC skew when existing.

According to another embodiment of the invention, can realize fabulous radio communication device, wherein, can remove DC skew, and become accurately estimated frequency skew under the situation of DC skew, IQ imbalance and frequency shift (FS) can in the OFDM symbol that receives, exist simultaneously the time.

Radio communication device according to the embodiment of the invention uses direct converting structure to receive ofdm signal.Even ofdm signal comprises the DC skew, radio communication device also can be offset the Frequency offset estimation of carrying out high-speed, high precision by using difference filter to remove DC.In addition, if owing to the gain of low noise amplifier changes the quick change that produces the DC skew, then from the output of difference filter, detect the DC skew and change, and Frequency offset estimation is not carried out in this output.Therefore, can increase the precision of Frequency offset estimation.

In addition, according to embodiments of the invention, in the time will comprising the vector signal addition of frequency offset information, vector signal be multiply by weighted factor according to level (that is the gain of low noise amplifier) corresponding to preamble.Therefore, during Frequency offset estimation, reduce IQ uneven and the time become in the DC bias effect, can use simple signal processing to carry out Frequency offset estimation more accurately.

By following the detailed description and the accompanying drawings to the preferred embodiment of the present invention, other features and advantages of the present invention will become apparent.

Description of drawings

Fig. 1 is the diagrammatic sketch that illustrates according to the receiver structure in the radio communication device of the embodiment of the invention;

Fig. 2 is the diagrammatic sketch that is illustrated in the subcarricr structure of the OFDM symbol in the Wireless LAN system that meets IEEE 802.11a/g standard;

Fig. 3 illustrates the diagrammatic sketch that ratio when DC offset power and ofdm signal power is the square error of 30dB and the estimated frequency deviant that obtains during by the standardized frequency offseting value of subcarrier interval;

Fig. 4 is the diagrammatic sketch that another receiver structure example in the radio communication device is shown;

Fig. 5 is the diagrammatic sketch that another receiver structure example in the radio communication device is shown;

Fig. 6 is the diagrammatic sketch that another receiver structure example in the radio communication device is shown;

Fig. 7 is the diagrammatic sketch that another receiver structure example in the radio communication device is shown;

Fig. 8 illustrates the diagrammatic sketch that the structure example of Frequency offset estimation and correction, bag detection and the thick peripheral synchronous circuit that regularly detects is carried out in the output of using high pass filter;

Fig. 9 illustrates the diagrammatic sketch that the structure example of Frequency offset estimation and correction, bag detection and thick another the peripheral synchronous circuit that regularly detects is carried out in the output of using high pass filter;

Figure 10 illustrates the diagrammatic sketch that is used for regulating by on-off controller 28 the method example of switching timing;

Figure 11 A is the diagrammatic sketch that the structure example of the high pass filter 21 that uses difference filter is shown;

Figure 11 B illustrates the diagrammatic sketch of use based on another structure example of the high pass filter 21 of the DC offset estimator of rolling average and adjuster;

Figure 12 is the diagrammatic sketch that the structure example of DC offset estimator 25 is shown;

Figure 13 is the diagrammatic sketch that the structure example of the peripheral synchronous circuit that is included in the circuit module of carrying out frequency offset correction and channel estimating among the LTS is shown;

Figure 14 illustrates to be configured so that the diagrammatic sketch of structure example of peripheral synchronous circuit of the frequency shift (FS) of the receiving baseband signal part after the LTS is estimated and removed to frequency offset estimator 22 and frequency offset corrector 24 respectively;

Figure 15 is the diagrammatic sketch that is illustrated in the preamble structure therefor of appointment among the IEEE 802.11a/g;

Figure 16 illustrates the schematic diagram that uses HPF to remove the receiver structure of DC skew;

Figure 17 A illustrates the diagrammatic sketch that HPF that use has an abundant small frequency with respect to subcarrier interval removes the DC component of ofdm signal;

Figure 17 B illustrates use to remove the DC component of ofdm signal with the diagrammatic sketch by nearly DC signal with respect to the HPF that subcarrier interval has big frequency;

Figure 18 is the schematic diagram that the receiver structure of estimating DC skew and frequency shift (FS) simultaneously is shown;

Figure 19 is the schematic diagram that the receiver structure of the parallel DC of estimation skew and frequency shift (FS) is shown;

Figure 20 is the schematic diagram that the receiver structure of repetition DC skew estimation and frequency offset compensation is shown;

Figure 21 is the diagrammatic sketch of instantiation that the structure of difference filter 5 and frequency offset estimator 6 is shown;

Figure 22 is the diagrammatic sketch that the output of the gain effects difference filter 5 by automatic gain control break low noise amplifier 2 is shown;

Figure 23 is the diagrammatic sketch that the uneven reason of IQ is shown;

Figure 24 is the diagrammatic sketch that the vector representation of the frequency offset information in the complex space is shown;

Figure 25 is the diagrammatic sketch that the instantiation of difference filter 5 and frequency offset estimator 6 structures is shown;

Figure 26 is illustrated in to gain in the low noise amplifier 2 to change before and the diagrammatic sketch of the estimate vector of frequency shift (FS) afterwards;

Figure 27 illustrates by the output with memory element 310 to be added in proper order corresponding to short preamble symbol t ₃And t ₄15 of multiplier 305 sampling output and corresponding to short preamble symbol t ₅And t ₁₀The diagrammatic sketch of 79 of the multiplier 305 resulting estimation composite vectors of sampling output;

Figure 28 be illustrate by according to the absolute value of the output signal of multiplier 305 with frequency shift (FS) with the weighted factor weighting and the frequency shift (FS) of weighting is added the diagrammatic sketch of the estimation composite vector that obtains mutually;

Figure 29 is the diagrammatic sketch that the Frequency offset estimation accuracy value (mean square error is to standardized frequency offseting value) of using the related art method and the method that proposes is shown;

Figure 30 is the configuration diagrammatic sketch that illustrates as the subcarrier of appointment in IEEE 802.11a/g;

Figure 31 illustrates the diagrammatic sketch of DC skew to the influence of frequency shift (FS);

Figure 32 illustrates to use difference filter 5 to remove the diagrammatic sketch of residual DC skew;

Figure 33 is the diagrammatic sketch that is illustrated in the structure of the receiver in the radio communication device in accordance with another embodiment of the present invention;

Figure 34 is the diagrammatic sketch of instantiation that the structure of the uneven estimator 1000 of difference filter 5 in the receiver shown in Figure 33, frequency offset estimator 6 and IQ is shown;

Figure 35 is the diagrammatic sketch of the MSE during the α in the gain that the is illustrated in LNA environment that do not have to change estimates, wherein, and α=0.05 and θ=5 °;

Figure 36 is the diagrammatic sketch of the MSE during the θ in the gain that the is illustrated in LNA environment that do not have to change estimates, wherein, and α=0.05 and θ=5 °

Figure 37 is the diagrammatic sketch of instantiation that the structure of difference filter 5, frequency offset estimator 6 and the uneven estimator 1000 of IQ is shown;

Figure 38 is the diagrammatic sketch that the structure example of another receiver in the radio communication device is shown;

Figure 39 is the diagrammatic sketch that the structure example of another receiver in the radio communication device is shown;

Figure 40 is the diagrammatic sketch that the structure example of another receiver in the radio communication device is shown;

Figure 41 is the diagrammatic sketch that the structure example of another receiver in the radio communication device is shown;

Figure 42 illustrates the diagrammatic sketch that the structure example of Frequency offset estimation and correction, bag detection and the thick peripheral synchronous circuit that regularly detects is carried out in the output of using high pass filter;

Figure 43 illustrates the diagrammatic sketch that the structure example of Frequency offset estimation and correction, bag detection and thick another the peripheral synchronous circuit that regularly detects is carried out in the output of using high pass filter;

Figure 44 is the diagrammatic sketch that the structure example of the peripheral synchronous circuit that is included in the circuit module of carrying out frequency offset correction, IQ disequilibrium regulating and channel estimating among the LTS is shown; And

Figure 45 is the diagrammatic sketch that the structure example of the peripheral synchronous circuit that is included in the circuit module of carrying out frequency offset correction, IQ disequilibrium regulating and channel estimating among the LTS is shown.

Embodiment

First embodiment

Describe the first embodiment of the present invention in detail below with reference to accompanying drawing.

The present invention designs a kind of radio communication device that is used to use direct converting structure reception ofdm signal.Do not use the direct converting structure of IF filter easily to realize broadband receiver, and increased the flexibility of receiver design.

The problem of ofdm communication system is that the little error between the frequency of oscillator in the transmitter and receiver can cause frequency shift (FS), and it is counted as the phase place rotation phenomenon of the received signal in the numerical portion of receiver.In common program, use the known training sequence in the packet header of adding each bag to come Observed Frequency Offset, and correcting frequency shift.

Yet, directly change receiver and have the problem that causes DC component or DC skew owing to the self-mixing of local signal in output place of low-converter.Frequency offset estimation and the precision that regularly detects are easy to be subjected to the influence of DC skew, are difficult to accurately estimated frequency skew under the situation that has the DC skew.

Radio communication device according to the embodiment of the invention has been realized at a high speed and the high accuracy Frequency offset estimation by using difference filter to remove the DC skew.

Fig. 1 shows the structure according to the receiver in the radio communication device of first embodiment of the invention.Device shown in Figure 1 has the module that is used to receive the direct conversion receiver of ofdm signal and is used for the compensating frequency skew.

When antenna receives ofdm signal, have only that the signal of desired frequency band is transmitted through band pass filter (BPF) 1 in the ofdm signal, and amplified by low noise amplifier (LNA) 2.The RF signal that receives has the frequency shift (FS) that is caused by the frequency error between the local oscillator of transmitter and receiver.

The gain of automatic gain control (AGC) circuit adjustment low noise amplifier 2 maintains suitable constant level with the power with received signal.For example, in IEEE 802.11a/g, specify 50dB or bigger gain control range.Usually, when input begins, big gain is set in low noise amplifier 2, then, for example around the center of short preamble period (in first embodiment, at the 4th short preamble t ₄The end), switch to than low gain according to the power of received signal.Gain switch level (gain switchinglevel) is about 20dB.AGC mechanism is well-known, repeats no more herein.

Use frequency mixer 3 received signal of amplifying to be multiply by the local frequency f that produces by local oscillator 11 _L0, and use direct translative mode that its frequency inverted is baseband signal.Convert baseband signal to digital signal by modulus (AD) transducer (ADC) 4.

In the direct converting structure of receiver, because receive frequency and local frequency are equal, so cause DC component or DC skew in output place of low-converter by the self-mixing of local signal.If the gain of low noise amplifier 2 is changed by automatic gain control, then the DC skew (for example also changes in time, referring to IEEE 802.11a, part 11: WLAN medium accesses control (MAC) layer and physical layer (PHY) illustrate: the high-speed physical layer in the 2.4GHZ frequency band).When having, the receiving baseband signal in the prime becomes DC skew and frequency shift (FS).

The predetermined period of the preamble in the digital baseband signal is divided and inputs to difference filter 5.After removing the DC offset component, the signal that obtains is inputed to frequency offset estimator 6, to estimate more accurate frequency shift (FS) the signal of removing the DC skew from being offset by the DC that deducts estimation.Figure 21 shows the instantiation of the structure of difference filter 5 and frequency offset estimator 6.Difference filter 5 comprises delay cell 201 and adder 202.Frequency offset estimator 6 comprises delay cell 203, complex conjugate counting circuit 204, multiplier 205, adder 206, memory element 207 and phase detecting circuit 208.

Difference filter 5 is a kind of high pass filters with simple circuit structure and high response.When lacking preamble t ₄End when changing into the gain of low noise amplifier 2 than low gain (as mentioned above) according to the power of received signal, the DC offset level changes.The change of 5 pairs of DC offset level of difference filter has very high response, thereby prevents that high fdrequency component from therefrom passing through.

Be used for the cut-off frequency f that the DC skew is removed if increase _c, then the response to the change of the DC that gain the produced skew by changing low noise amplifier 2 increases.Yet, even also can end nearly DC signal, cause the problem (referring to Figure 17 B) of demodulation characteristics deterioration.On the contrary, dispose receiver shown in Figure 1, make and divide the preamble portion (that is the STS that, has big relatively subcarrier interval) that is used for Frequency offset estimation for DC is offset to remove.In other words, only the preamble portion of carrying out Frequency offset estimation is removed the DC skew, and the preamble portion after the part that is used for Frequency offset estimation and payload portions are removed frequency shift (FS) but do not remove the DC skew.Therefore, even cut-off frequency f _cIncrease, also can not produce adverse influence the demodulation characteristics of follow-up data division (after having the LTS of short subcarrier interval).

If DC skew then can be by difference filter 5 transmission DC skews because gain change of low noise amplifier 2 etc. and changing fast.When this impulse waveform is inputed to frequency offset estimator 6, may increase mean square error (MSE).Therefore, in case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6.When the detection signal of input is when comprising the signal of DC skew, frequency offset estimator 6 is determined from the samplings of difference filter 5 outputs, and can not carried out Frequency offset estimation to this sampling output.Therefore, can keep high estimated accuracy.

For example, the level of number of times that can change based on gain and received signal calculates the threshold value that difference filter 5 is used to detect the quick change of DC skew.

The estimated frequency shift value is input to frequency offset corrector 7, and the frequency shift (FS) of the baseband signal of the OFDM symbolic component of compensation after being used for the part of Frequency offset estimation.

The output of frequency offset corrector 7 is input to discrete Fourier transform (DFT) (DFT) unit 8, and demodulation is arranged in the subcarrier signal in the frequency domain.

The ofdm signal of supposing to input to according to the receiver of first embodiment of the invention does not comprise DC subcarrier (DC is corresponding to the 0Hz in the baseband signal in the OFDM demodulation).Frequency offset estimator 6 is estimated the preamble (that is, in having the STS of relatively large subcarrier interval) of each bag and is had frequency shift (FS) in the preamble of the identical ofdm signal symbol of twice transmission.

Wireless communication system according to the embodiment of the invention is the Wireless LAN system that meets IEEE 802.11a/g standard.Fig. 2 shows the subcarricr structure of the OFDM symbol in this Wireless LAN system.As shown in Figure 2, an OFDM symbol is made of 64 subcarriers, and 52 subcarriers wherein are modulated into information signal, and 4 subcarriers are as pilot signal.Transmission signals (that is, residual subcarrier carries spacing wave) not on the residual subcarrier that comprises the DC component.

When will be after being used for the part of Frequency offset estimation (promptly, after LTS) receiving baseband signal part (it is subjected to the influence of DC skew and frequency shift (FS)) when inputing to frequency offset corrector 7, accurately compensation and frequency, demodulation frequency skew is not removed the deterioration that DC is offset the demodulation characteristics that causes and can not bring by high pass filter.

Figure 15 shows the preamble structure therefor of appointment in IEEE 802.11a/g.In long preamble period, the OFDM symbol that is made of long training sequence (LTS) symbol of 3.2 microseconds is transmitted twice continuously.I sampling of the time waveform of OFDM symbol represented by s (i).Sampling s (0), s (1) ..., s (63) } relevant with an OFDM symbol, and sampling { s (64), s (65), ..., s (127) } relevant with the 2nd OFDM symbol (if the rank of discrete Fourier transform (DFT) are represented that by N then an OFDM symbol is sampling { s (0), s (1) ..., s (N/4-1) set, and the 2nd the OFDM symbol be the sampling { s (N/4), s (N/4+1) ..., s (2N/4-1) set).

If the frequency shift (FS) of this moment represented by Δ f, DC skew represented by D, and then relevant with i short preamble receiving baseband signal is provided by following equation:

r(i)＝s(i)exp(j2πΔfi)+D …(7)

Difference filter 5 comprises delay cell 201 and adder 202.The AD switching signal that obtains by AD converter 4 is input to and postpones 201 input.The AD switching signal also is input to the first input end of adder 202, and the output of delay cell 201 is input to second input of adder 202 then through conversion, is used for subtracting each other between them.Therefore, difference filter 5 is handled received baseband signal according to following equation (8):

d(i)＝r(i+1)-r(i)

＝s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi) …(8)

Equation (8) expression is with respect to the output signal of the difference filter 5 of i short preamble.

Frequency offset estimator 6 in the level of back comprises delay cell 203, complex conjugate counting circuit 204, multiplier 205, adder 206, memory element 207 and phase detecting circuit 208.The output of adder 202 is input to the first input end of delay cell 203 and multiplier 205.Delay cell 203 postpones the individual sampling of N/4 (=16) corresponding to short preamble length with input signal, and the signal that postpones is inputed to complex conjugate counting circuit 204 in the level of back.The output of complex conjugate counting circuit 204 is input to second input of multiplier 205.Therefore, multiplier 205 is to each short preamble t ₁, t ₂Deng carrying out the cross-correlation operation that provides by following equation.

$d(i+16)/d\left(i\right)=\frac{s(i+1+16)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+1+16)\right)-s(i+n)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+16)\right)}{s(i+1)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}(i+1)\right)-s\left(i\right)\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;f}\left(i\right)\right)}$

$=\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\left(16\right)\right)\&CenterDot;\&CenterDot;\&CenterDot;\left(9\right)$

The output of multiplier 205 is connected to first end of adder 206, and the output of memory element 207 is connected to second input of adder 206.The output of adder 206 is input to memory element 207 and phase detecting circuit 208.Then, use adder 206 to lack the cross correlation results addition that preambles are determined by above equation (9) to all, and estimated frequency shifted by delta f.

The frequency shift (FS) of the part (being subjected to the influence of frequency shift (FS)) of the receiving baseband signal of frequency of utilization offset corrector 7 compensation after being used as the part of Frequency offset estimation.Particularly, come correcting frequency shift according to frequency displacement by the reversal data phase place.Short training sequence (STS) can also be used for to carry out Frequency offset estimation (in this case, hits is 16) with above similar mode.

Fig. 3 shows the mean square error that ratio when DC offset power and ofdm signal power is 30dB and the estimated frequency deviant that obtains during by the standardized frequency offseting value of subcarrier interval.As can be seen from Figure 3, realize accurate Frequency offset estimation by using difference filter 5 to be offset from the reception preamble removal DC that is used for Frequency offset estimation.Because STS has the repetition (referring to Figure 15) of identical training sequence symbols, so can use the skew of similar operations estimated frequency.

In the superincumbent equation (9), suppose DC shift constant or not conversion in time, and difference filter 5 can be deleted the DC skew.Yet when changing the gain of low noise amplifier 2, the change of DC offset level occurs as high fdrequency component, and the amplitude of DC skew is represented in the output of difference filter 5.

Particularly, automatic gain control circuit is that low noise amplifier 2 is provided with big gain when input begins, and uses first to fourth short preamble t ₁To t ₄Be identified for the power of received signal is maintained suitable gain (because the influence of multichannel and do not use preamble t of constant level ₁And t ₂).At the 5th short preamble t ₅Begin gain is switched to than low gain.The gain switch level is about 20dB.The DC offset level also according to the gain switching and conversion, this is at preamble t ₅Influence the output (referring to Figure 22) of difference filter 5 during beginning.

If the DC deviant of i sampling is by D (i) expression, then when because the gain change of low noise amplifier 2 etc. when causing DC in the OFDM sampling to be offset change, are determined the output of difference filter 5 by following equation (10):

d(i)＝r(i+1)-r(i)

＝{s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi)}+{D(i+1)-D(i)}

…(10)

Understand from above equation, when the DC skew changes fast, poor (D (i+1)-D (i)) between the DC skew of residual (i+1) individual sampling and the DC skew of i sampling, and the absolute value of the output d (i) of difference filter 5 increases.Therefore, when detecting DC skew change, do not remove the DC skew and pass through difference filter 5 transmission.

In receiver shown in Figure 1, surpass predetermined value in case detect the absolute value of output d (i), difference filter just provides indication (detection signal) to the frequency offset estimator 6 in the level of back, thereby i the sampling output d (i) that comprises the DC bias effect is not carried out Frequency offset estimation.Thereby frequency offset estimator 6 is the estimated frequency skew under the situation of not considering the DC skew, thereby has improved estimated accuracy.Even under the situation that the frequent automatic gain of carrying out low noise amplifier 2 is controlled, also the DC offset component of transmission is not carried out Frequency offset estimation, thereby keeps high estimated accuracy.

As mentioned above, owing to,, handle the situation that DC skew and frequency shift (FS) occur simultaneously so the structure of meeting sb. at the airport shown in Fig. 1 can realize accurate Frequency offset estimation using difference filter 5 to remove DC skew estimated frequency skew afterwards.In addition, owing to got rid of the estimated frequency deviant that in the cycle that changes low noise amplifier 2 gains, obtains, so can keep high estimated accuracy.

In receiver structure shown in Figure 1, remove by frequency offset estimator 6 estimated frequency shift in the receiving baseband signal part (that is, therefrom not removing the DC skew) of frequency offset corrector 7 after the part that is used for Frequency offset estimation.In other words, when the preamble portion of estimated frequency skew is removed the DC skew, and the preamble portion after the part that is used for Frequency offset estimation and payload portions are removed frequency shift (FS) but do not remove the DC skew.In this case, even may produce the problem that the very high but DC of Frequency offset estimation precision skew also can influence the demodulator circuit in the level of back.

Fig. 4 shows the example of the receiver circuit that is used to address this problem.In the structure of receiver shown in Figure 1, after from the receiving baseband signal that inputs to difference filter 5, removing the DC skew, use by frequency offset estimator 6 estimated frequency shift, make the DC skew of the receiving baseband signal after being used for the part of Frequency offset estimation comprise that part stands frequency offset compensation by frequency offset corrector 7.On the other hand, in the structure of receiver shown in Figure 4, after removing the DC skew, based on passing through the phase place of frequency offset estimator 6 estimated frequency shift counter-rotating by the local frequency of local oscillator 11 vibrations by difference filter 5.Therefore, can comprise the influence of partly removing DC skew and frequency shift (FS) simultaneously from the DC skew of the receiving baseband signal after the part that is used for Frequency offset estimation.

Equally, in this case, in case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6.As mentioned above, when the input detection signal, frequency offset estimator 6 is not carried out Frequency offset estimation to the sampling output that obtains from difference filter 5, thereby avoids the influence by the DC skew of difference filter 5 transmission.

Fig. 5 be can be after the part that is used for Frequency offset estimation the receiving baseband signal part of (after the LTS) remove the structure example of another receiver of DC bias effect.

In receiver shown in Figure 5, the predetermined period of preamble is divided and inputs to difference filter 5 with removal DC skew in the receiving baseband signal, and the skew of frequency offset estimator 6 estimated frequencies.In case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6.As mentioned above, when the input detection signal, frequency offset estimator 6 is not carried out Frequency offset estimation to the sampling output that obtains from difference filter 5, to avoid the influence by the DC skew of difference filter 5 transmission.

Be parallel to Frequency offset estimation and handle, DC offset estimator 9 is estimated the DC skew of receiving baseband signal, and DC offset corrector 10 is removed the DC skew from receiving baseband signal.Then, based on the high accuracy frequency offseting value of estimating after removing the DC skew, 7 pairs of frequency offset corrector have been removed the part of the part that the is used for Frequency offset estimation receiving baseband signal afterwards of DC skew and have been carried out frequency offset compensation.

Usually, by being asked, the digital baseband signal of conversion on average estimates the DC skew.Therefore, if DC skew because gain change of low noise amplifier 2 etc. and changing fast, then mean value is die on, and the DC skew estimate will cause the precision deterioration continuously.Therefore, in case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6 and DC offset estimator 9.DC offset estimator 9 is got rid of the data estimator of estimating before the input detection signal, reappraise the DC skew then, to prevent the reduction of estimated accuracy.

Fig. 6 show can be after the part that is used for Frequency offset estimation the part of the receiving baseband signal of (after the LTS) remove the structure example of another receiver of DC bias effect.

In receiver shown in Figure 6, frequency mixer 3 will be by receiving the local frequency f of this signal times to produce by local oscillator 11 _L0, use direct converting structure that this received signal is carried out down-conversion.Final receiving baseband signal comprises by caused DC skews (referring to the part (a) of Figure 32) such as the self-mixings of local signal.By DC is offset by high pass filter (HPF) 12 its removal.Because high pass filter 12 allows all signals to pass through, thus with the cut-off frequency of high pass filter 12 be provided with relatively low, make the nearly DC signal in the OFDM symbol.To convert digital signal to by the baseband signal of high pass filter 12 transmission by AD converter (ADC) 4.

Has low cut-off frequency f _c High pass filter 12 guarantee good demodulation characteristics in the back level, but the change of DC skew is had low-response.Therefore, if change owing to the DC skew takes place for the gain switching of low noise amplifier 2 etc., then the influence of Gai Bianing can keep for a long time, and the DC skew is transmitted continuously and passed through high pass filter 12 (referring to the part (b) of Figure 32).In order to handle this situation, the predetermined period of the preamble in the receiving baseband signal is divided into two branches, branch be input to have higher cut off frequency difference filter 5 to remove residual DC skew.Shown in the part (c) of Figure 32, difference filter stops residual DC skew, and at preamble t ₅Begin locate to change gain and the time only export strong impulse waveform.

Then, frequency offset estimator 6 comes the estimated frequency skew based on the autocorrelation value of difference filter 5 outputs.Frequency offset corrector 7 is removed frequency shift (FS) from receiving baseband signal.

In case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6.If impulse waveform is input to frequency offset estimator 6, then MSE may increase.Therefore, as mentioned above, when the input detection signal, frequency offset estimator 6 is not carried out Frequency offset estimation to the sampling output that obtains from difference filter 5, to avoid the influence by the DC skew of difference filter 5 transmission.

As mentioned above, number of times that can change based on gain and received signal level calculate the threshold value that difference filter 5 is used to detect the quick change of DC skew.For example, received signal intensity indication (RSSI) circuit can be arranged in the back level of high pass filter 12, to detect received signal level.

DC offset estimator 9 estimates to be used for the DC skew of the part of the receiving baseband signal after the part of Frequency offset estimation, and DC offset corrector 10 is removed DC and is offset from receiving baseband signal.

In case receive the quick change of DC skew, difference filter 5 just inputs to detection signal DC offset estimator 9.As mentioned above, DC offset estimator 9 is got rid of the data estimator of estimating before the input detection signal, and reappraises the DC skew, to prevent the reduction of estimated accuracy.

Then, based on the high accuracy frequency offseting value of estimating after removing the DC skew, 7 pairs of frequency offset corrector are used for the part of part (the having removed the DC skew) receiving baseband signal afterwards of Frequency offset estimation and carry out frequency offset compensation.

Fig. 7 show can be after the part that is used for Frequency offset estimation the part of the receiving baseband signal of (after the LTS) remove the structure example of another receiver of the influence of DC skew.

In receiver shown in Figure 7, frequency mixer 3 is by multiply by received signal the local frequency f that is produced by local oscillator 11 _L0, use direct converting structure that this received signal is carried out down-conversion.Receiving baseband signal by 12 transmission of the high pass filter in the level of back obtain passes through the caused DC skews such as self-mixing of local signal with removal from receiving baseband signal.Because all signals that high pass filter 12 allows to be included in the bag pass through, so the cut-off frequency f of high pass filter 12 _cWhat be provided with is relatively low, and making can be by the nearly DC signal in the OFDM symbol.Then, convert receiving baseband signal to digital signal by AD converter (ADC) 4.

Has low cut-off frequency f _c High pass filter 12 guarantee good demodulation characteristics in the back level, but the change of DC skew is had low-response.Therefore, if change owing to the DC skew takes place for the gain switching of low noise amplifier 2 etc., then the influence of Gai Bianing can keep for a long time, and the DC skew continuously transmission pass through high pass filter 12 (referring to the part (b) of Figure 32).In order to solve this situation, the predetermined period of the preamble in the receiving baseband signal is divided into two branches, with a branch input to have higher cutoff frequency more difference filter 5 to remove residual DC skew.Shown in the part (c) of Figure 32, difference filter 5 stops residual DC skew, and at preamble t ₅Begin locate to change gain and the time only export strong impulse waveform.

Based on the phase place of estimated frequency shift counter-rotating by the local frequency of local oscillator 11 vibrations.Therefore, can comprise the influence of partly removing DC skew and frequency shift (FS) simultaneously from the DC skew of the receiving baseband signal after the part that is used for Frequency offset estimation.

In case detect the quick change of DC skew, difference filter 5 just inputs to detection signal frequency offset estimator 6.If impulse waveform is inputed to frequency offset estimator 6, then MSE increases.Therefore, as mentioned above, when the input detection signal, frequency offset estimator 6 is not carried out Frequency offset estimation to the sampling output that obtains from difference filter 5, to avoid the influence by the DC skew of difference filter 5 transmission.

Number of times that can change based on gain and received signal level calculate the threshold value that difference filter 5 is used to detect the quick change of DC skew.As mentioned above, for example, the RSSI circuit can be arranged in the back level of high pass filter 12 to detect received signal level.

In the first embodiment of the present invention, use the preamble estimated frequency skew that repeats to transmit twice identical ofdm signal symbol.In the preamble structure therefor of Figure 15, use the STS execution to be used to wrap the signal processing of detection, thick regularly detection and Frequency offset estimation and correction.In this signal processing, deterioration in characteristics is very responsive to the DC skew.

In receiver shown in Figure 6, by using the influence that removes the DC skew as a kind of output signal of difference filter 5 of high pass filter, to guarantee high-precision Frequency offset estimation.Can also use the output signal of high pass filter to carry out the bag detection and slightly regularly detect, to prevent being offset caused deterioration in characteristics owing to DC.

Fig. 8 shows the structure example that Frequency offset estimation and correction, bag detection and the thick peripheral synchronous circuit that regularly detects are carried out in the output of using high pass filter.

Synchronous circuit shown in Figure 8 comprises about the guiding high pass filter 21 of each I axle and Q axle input signal and the path of DC offset estimator 25, and open or close two switches 26 and 27 specially to switch between two path.

STS is in the relatively large subcarrier interval that has 1.25MHz between the subcarrier near DC.Given this, with the cut-off frequency of high pass filter 21 be provided with higher, with the Expected Response characteristic of guaranteeing that skew changes to DC.Therefore, in minimum SNR loss, can suppress the influence of DC skew.

During Synchronous Processing, DC offset estimator 25 is estimated the DC skew from receiving baseband signal.STS has the cycle of 0.8 μ s, and can use rolling average to wait and estimate the DC skew.If preceding four symbol t of STS ₁To t ₄Be used to automatic gain control, DC migration processing etc., then synchronous circuit can obtain six symbol t the biglyyest ₅To t ₁₀Estimated time, that is, and 4.8 microseconds, and guarantee that high-precision DC skew estimates.

During the cycle before the STS end, the IQ input is connected to the path of guiding high pass filter 21, and frequency offset estimator 22 and bag detects and thick timing detector 23 carries out with high accuracy that Frequency offset estimation and bag detect respectively from the receiving baseband signal of removing the DC skew and slightly regularly detect.

At the end of STS, change over switch 26 and 27 open/close state, and the IQ input is switched to the path of guiding DC offset corrector from the path of guiding high pass filter 21.Use the output signal of thick timing detector 23 in the end of STS.

DC offset estimator 25 uses the before sufficiently long time of the end of STS to carry out DC skew estimation, thereby realizes high-precision DC offset correction.After the STS end, there is not the path of the high pass filter 21 that leads.That is, the part of not using 21 pairs of high pass filters with higher cutoff frequency to have the LTS receiving baseband signal afterwards of short subcarrier interval is carried out the DC skew and is removed, and does not consider the deterioration among the SNR.

In addition, during the cycle before the STS end, frequency offset corrector 24 is used by frequency offset estimator 22 estimated frequency shift, and the part of the receiving baseband signal after the STS end is carried out frequency offset correction.

In the structure of synchronous circuit shown in Figure 8, the switching time between the path of the definite guiding of the thick regularly detection signal of use high pass filter 21 and the path of the frequency offset corrector that leads.In view of the processing delay of digital circuit, can be before the STS end toggle path.

In the structure of synchronous circuit shown in Figure 9, the detection signal in response to thick timing detector 23 does not come direct diverter switch 26 and 27, but is provided for control switch 26 and 27 on-off controllers 28 that switch in addition.Figure 10 shows and is used to use on-off controller 28 to regulate the method example of switching timing.

On-off controller 28 uses the moving average of the correlation of input signal to determine switching timing.Switch controller 28 is provided with predetermined threshold, and will export switch 26 and 27 as the timing of control signal to when moving average surpasses threshold value.As shown in figure 10, by threshold value is set, can be in the cycle that detects thick timing from bag the flexible switching timing.

Figure 11 A and Figure 11 B show the structure example of high pass filter 21.Figure 11 A shows the structure of using difference filter, and Figure 11 B shows use based on the DC offset estimator of rolling average and the structure of adjuster.

Figure 12 shows the structure example of DC offset estimator 25.In structure shown in Figure 12, the estimation DC deviant that the switching timing that can use the output signal of on-off controller 28 or thick timing detector 23 to remain on switch 26 and 27 obtains.Therefore, the DC offset corrector can be used the high accuracy DC deviant of estimating in the STS end, and the part of the receiving baseband signal after the STS end is carried out the DC offset correction.

Fig. 8 and synchronous circuit shown in Figure 9 are configured to only to use the STS in the preamble structure therefor of appointment in IEEE 802.11a, consider that DC is offset to carry out frequency offset correction, and can use the LTS after the STS end arbitrarily.

Usually, after using STS execution coarse frequency offset correction, use LTS to carry out thin frequency offset correction and channel estimating.Therefore, in STS, estimate accurate frequency shift (FS), to realize that more precise channels is estimated.

Figure 13 shows the structure example of the peripheral synchronous circuit that is included in the circuit module of carrying out frequency offset correction and channel estimating in the long preamble period.In example shown in Figure 13, be provided for frequency offset estimator 31 and the frequency offset corrector 32 of LTS independently with frequency offset estimator 22 that is used for STS and frequency offset corrector 24, and in the level of back, channel estimator 33 be set.Frequency offset estimator 22 uses short preamble to carry out coarse frequency skew estimation (wherein, because the first and second short preamble t ₁And t ₂May be subjected to the influence of multichannel, so the 3rd and following short preamble t ₃To t ₁₀Be used to estimate).Frequency offset estimator 31 uses long preamble T ₁And T ₂Carry out thin Frequency offset estimation.

At the end of short preamble period, the IQ input switches to the path of guiding DC offset corrector from the path of guiding high pass filter 21.After LTS, estimated accurate DC of sufficiently long time was offset before each DC offset corrector deducted from each IQ input signal and uses the STS end, to proofread and correct the DC skew.Then, frequency offset corrector 24 is proofreaied and correct during the cycle before the STS end by frequency offset estimator 22 estimated frequency shift.

In case receive the part of removing the DC skew of using the STS estimation and the LTS receiving baseband signal afterwards that is offset, frequency offset estimator 31 is just carried out thin Frequency offset estimation.Remove by frequency offset estimator 31 estimated frequency shift the part of the receiving baseband signal of frequency offset corrector 32 after LTS.

Channel estimator 33 uses the receiving baseband signal that has used LTS to remove less important (residual) frequency shift (FS) to carry out channel estimating with higher precision.

In circuit structure shown in Figure 13, the circuit module of the part of the receiving baseband signal after the LTS being carried out Frequency offset estimation and removal is set additionally.Alternatively, frequency offset estimator 22 and frequency offset corrector 24 can be configured to estimate respectively and remove the frequency shift (FS) in the part of the receiving baseband signal after the LTS.Figure 14 shows the structure example of the peripheral synchronous circuit under the kind situation of back.

At the end of STS, the IQ input switches to the path of guiding DC offset corrector from the path of guiding high pass filter 21.After LTS, each DC offset corrector deducts the estimated accurate DC of terminal preceding sufficiently long time that uses STS and is offset to proofread and correct the DC skew from each IQ input signal.Then, frequency offset corrector 24 is proofreaied and correct during the cycle before the STS end by frequency offset estimator 22 estimated frequency shift.

In addition, in the end of STS, open switch 27 is used for the output of frequency offset corrector 24 is turned back to frequency offset estimator 22 with generation path.

Frequency offset estimator 22 receives the part of having removed the receiving baseband signal after the LTS that uses DC skew that STS estimates and frequency shift (FS), and further estimated frequency is offset.Remove by frequency offset estimator 22 estimated frequency shift the part of the receiving baseband signal of frequency offset corrector 24 after LTS.

Channel estimator 33 uses and has used receiving baseband signal that LTS removes less important (residual) frequency shift (FS) to carry out channel estimating with high accuracy more.

Described a kind of directly the conversion and be used for when DC skew changes according to the gain of low noise amplifier, reducing the method that Frequency offset estimation is handled the influence that change DC is offset in the receiver at OFDM.

OFDM directly changes the problem that receiver also has the IQ imbalance and is offset by the caused DC of the self-mixing of local signal.Cause the IQ imbalance by phase difference and the difference of vibration between the frequency mixer between the local signal that inputs to I axle and Q axle frequency mixer.Similar with the DC skew, the IQ imbalance causes the deterioration of Frequency offset estimation precision, and influences decoding characteristics.Below will be described in detail in exist IQ uneven and the time become the method that estimated frequency is offset under the situation that the DC skew exists.

In receiver structure shown in Figure 1, the DC offset level changes according to the gain of low noise amplifier 2.When the DC of variation skew is equal to or greater than predetermined value, to symbol execution Frequency offset estimation (referring to Figure 22) from difference filter 5 outputs, thus the influence of the DC skew of the variation in the processing of reduction Frequency offset estimation.Yet receiver shown in Figure 1 does not fully take into account the unbalanced influence of IQ.

OFDM directly changes the problem that receiver has the IQ imbalance and is offset by the caused DC of the self-mixing of local signal.Directly converting structure does not use the IF signal of numeric field, and the demodulation of IQ integration is not to carry out in numeric field but carry out in analog domain.Thereby, cause the IQ imbalance by the imbalance between homophase (I) component and quadrature phase (Q) component.Particularly, cause the IQ unbalance in phase by the non-90 degree phase differences between the local signal that inputs to I channel and Q channel mixer, and cause that by the gain inequality between the signal in I channel and the Q channel IQ gain is uneven.

Figure 23 shows the IQ reason of unbalanced.In Figure 23, will be divided into two branches from the local signal of single local oscillator, with 90 ° of the phase-shifts of a branch to generate cosine signal and sinusoidal signal.If two signals have the phase difference more than 90 ° or have different amplitudes, then can distortion through the baseband signal of frequency inverted.This phenomenon is called as the IQ imbalance.By the influence of difference filter transmission distortion, therefore, the IQ imbalance causes the deterioration of Frequency offset estimation precision.

If the phase difference between cosine signal and the sinusoidal signal is represented that by θ difference of vibration is represented (with decibel (dB) expression), then local frequency f by λ _cThe I component of local signal and Q component by following formulate, and be input to corresponding frequency mixer:

I component: (1+ α) cos (2 π f _cT-θ/2)

Q component :-(1-α) sin (2 π f _cT+ θ/2)

Wherein, α is represented by the following equation that uses difference of vibration λ:

$\mathrm{\&alpha;}=\frac{{10}^{\mathrm{\&lambda;}/20}-1}{{10}^{\mathrm{\&lambda;}/20}+1}$

Use corresponding frequency mixer that the component and received signal r (i) frequency of local signal are multiplied each other.Multiple transmission symbol is set to X _n=a _n+ jb _n

$r\left(t\right)(1+\mathrm{\&alpha;})\cos (2\mathrm{\&pi;}{f}_{c}t-\frac{\mathrm{\&theta;}}{2})$

$=\frac{(1+\mathrm{\&alpha;})}{2}\underset{n=-N/2}{\overset{N/2}{\mathrm{\&Sigma;}}}\left[(a_n-b_n)\right\{\cos (2\mathrm{\&pi;}(2f_c+n{f}_0)t-\frac{\mathrm{\&theta;}}{2})+\cos (2\mathrm{\&pi;n}{f}_{0}t+\frac{\mathrm{\&theta;}}{2})\}-$

$(a_n+b_n)\{\sin (2\mathrm{\&pi;}(2f_c+n{f}_0)t-\frac{\mathrm{\&theta;}}{2})+\sin (2\mathrm{\&pi;n}{f}_{0}t+\frac{\mathrm{\&theta;}}{2})\}]$

$r\left(t\right)\{-(1-\mathrm{\&alpha;})\sin (2\mathrm{\&pi;}{f}_{c}t+\frac{\mathrm{\&theta;}}{2})\}$

$=-\frac{(1-\mathrm{\&alpha;})}{2}\underset{n=-N/2}{\overset{N/2}{\mathrm{\&Sigma;}}}\left[(a_n-b_n)\right\{\sin (2\mathrm{\&pi;}(2f_c+n{f}_0)t+\frac{\mathrm{\&theta;}}{2})-\sin (2\mathrm{\&pi;n}{f}_{0}t-\frac{\mathrm{\&theta;}}{2})\}$

$+(a_n+b_n)\{\cos (2\mathrm{\&pi;}(2f_c+n{f}_0)t+\frac{\mathrm{\&theta;}}{2})-\cos (2\mathrm{\&pi;n}{f}_{0}t-\frac{\mathrm{\&theta;}}{2})\}]$

Do not consider that the unbalanced receiving baseband signal of IQ is definite by above equation (7), and the reception complex baseband signal under the uneven influence of IQ is determined by following equation (11):

$\hat{r}\left(i\right)=(\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{j\&alpha;}\sin \frac{\mathrm{\&theta;}}{2})r\left(i\right)+(\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}-j\sin \frac{\mathrm{\&theta;}}{2})r^{*}\left(i\right)\&CenterDot;\&CenterDot;\&CenterDot;\left(11\right)$

$\hat{r}\left(i\right)=\mathrm{\&phi;r}\left(i\right)+{\mathrm{\&psi;}}^{*}{r}^{*}\left(i\right)$

Wherein

$\mathrm{\&phi;}=\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{j\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}$

$\mathrm{\&psi;}=\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}+j\sin \frac{\mathrm{\&theta;}}{2}$

In equation (11), i is illustrated in the hits in the short preamble, and goes up asterisk (*) expression complex conjugate.

Therefore, the differential signal from difference filter 5 outputs under the unbalanced influence of IQ is determined by following equation (12):

$d\left(i\right)=\hat{r}(i+1)-\hat{r}\left(i\right)$

$=\mathrm{\&phi;}(r(i+1)-r\left(i\right))+{\mathrm{\&psi;}}^{*}(r^{*}(i+1)-r^{*}\left(i\right))\&CenterDot;\&CenterDot;\&CenterDot;\left(12\right)$

Frequency offset estimator 6 comes the estimated frequency skew based on the differential signal of being determined by above equation (12).Multiplier 205 multiply by the differential signal that postpones the N/4 sampling with differential signal, to obtain the Frequency offset estimation vector about each sampling.Comprise from the unbalanced autocorrelation value of IQ (that is Frequency offset estimation vector) of difference filter 5 outputs and representing by following equation (13):

$\hat{d}(i+\frac{N}{4}){\hat{d}}^{*}\left(i\right)$

$={\left|\mathrm{\&phi;}\right|}^2{|r(i+1)-r\left(i\right)|}^2\exp (\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;fN}/4)+{\mathrm{\&phi;}}^{*}{\mathrm{\&psi;}}^{*}{(r^{*}(i+1)-r^{*}\left(i\right))}^2\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;fN}/4)$

$+\mathrm{\&phi;\&psi;}{(r(i+1)-r\left(i\right))}^2\exp (\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;fN}/4)+{\left|\mathrm{\&psi;}\right|}^2{|r(i+1)-r\left(i\right)|}^2\exp (\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;fN}/4)\&CenterDot;\&CenterDot;\&CenterDot;\left(13\right)$

Frequency offset information by above equation (13) expression has four.As shown in figure 24, these are represented as the vector in the complex space.

First in the equation (13) is the vector that only depends on frequency shift (FS).That is, when the baseband signal that receives does not comprise that IQ is uneven, only from first of multiplier 205 output, and can estimated frequency skew from the angle of vector.

Second to the 4th in the equation (13) is that these cause the deterioration of Frequency offset estimation precision by the uneven item that produces of IQ.In the 4th, value | ψ | ²Little to can ignore (should be appreciated that and since α be about 0.1 and θ be about 0.05 °, so ψ is very little).Second and the 3rd is that complex conjugate is right, and these two phases are only produced real component, and it is counted as the main cause of Frequency offset estimation precision deterioration.

Multiplier 205 is determined the cross correlation results of each short preamble, uses adder 206 with the cross correlation results of all short preambles estimated frequency shifted by delta f in addition mutually.Be included in distortion in the summation of cross correlation results and depend on the sampling sum of synchronous signal mode.

Figure 21 shows the structure example of difference filter 5 and frequency offset estimator 6, wherein, though do not take into full account the unbalanced influence of IQ in the frequency offset estimation process, becomes the DC skew when having considered.Figure 25 shows and suppresses the uneven difference filter 5 that influences of IQ and the structure example of frequency offset estimator 6.In Figure 25, difference filter 5 comprises delay cell 301 and adder 302.Frequency offset estimator 6 comprises delay cell 303, complex conjugate counting circuit 304, multiplier 305 and 306, coefficient calculation circuit 307, adder 309, memory element 310 and phase detecting circuit 311.

Below with reference to Figure 25 describe by the IQ in the blanketing frequency skew estimation procedure uneven and the time become the operation of the more accurate Frequency offset estimation of influence execution of DC skew.

By the signal of band pass filter 1 transmission, and only amplify the ofdm signal of expectation by low noise amplifier 2 by the antenna reception.Use multiplier 3 that amplifying signal be multiply by local signal from local oscillator 11, and be converted into baseband signal.Convert the baseband signal that receives to digital signal by AD converter 4.If represented by α and θ respectively by difference of vibration and phase difference that the IQ imbalance causes, then the baseband signal of Jie Shouing is represented by above equation (11).

Difference filter 5 comprises delay cell 301 and adder 302.To input to the input of delay cell 301 by the AD switching signal that AD converter 4 obtains.Also the AD switching signal is inputed to the first input end of adder 302, and the output of delay cell 301 is used for subtracting each other between them through counter-rotating and be input to second input of adder 302.Therefore, difference filter 5 is handled the baseband signal that receives according to above equation (12).

Frequency offset estimator 6 in the level of back comprises delay cell 303, complex conjugate counting circuit 304, multiplier 305 and 306, coefficient calculation circuit 307, adder 309, memory element 310 and phase detecting circuit 311.The output of adder 302 is input to first end of delay cell 303 and multiplier 305.Delay cell 303 postpones the individual sampling of N/4 (=16) corresponding to short preamble length with input signal, and inhibit signal is inputed to complex conjugate counting circuit 304 in the level of back.The output of complex conjugate counting circuit 304 is input to second input of multiplier 305.Therefore, in frequency offset estimator 6, to short preamble t ₁, t ₂Deng in each carry out the operation that provides by above equation (13).

First in the equation (13) is the vector that comprises frequency shift (FS) Δ f.Use adder 309 and memory element 310 with addition of vectors, and detect the anglec of rotation of addition vector by phase detecting circuit 311.Therefore, the value of estimated frequency shifted by delta f.

As previously mentioned, second to the 4th in the equation (13) is by the uneven item that produces of IQ, and these cause the deterioration of Frequency offset estimation precision.As mentioned above, in the 4th, the absolute value of value ψ is little of ignoring.Second and the 3rd is that complex conjugate is right, and these two phases are only produced in addition will be as the real component of the main cause of Frequency offset estimation precision deterioration.

Second and the 3rd is the vector that depends on short preamble r (i) pattern and frequency shift (FS), and the direction of vector is unfixing.In the sampling of significant figure purpose, use memory element 310 and adder 309 will be from the estimated frequency deviant that is used for preamble symbol additions in fully many samplings of multiplier 305 outputs, thereby increase by first in the equation (13), simultaneously with respect to first distortion component that reduces in second to the 4th.Detect the phase place of vector by the phase detecting circuit in the level of back 311.Therefore, can estimate more accurate frequency shift (FS).

If the gain of low noise amplifier 2 changes, then be included in the increase of the IQ unbalanced component in the estimated frequency shift when big gain is set in the process of carrying out Frequency offset estimation by frequency offset estimator 6.Therefore, if with the simple addition of estimated frequency shift on a plurality of preamble symbol, then may be difficult to fully reduce the ratio of IQ unbalanced component.

Usually, in receiver, be that low noise amplifier 2 is determined big gain when commencing signal detects, after this, gain is become lower gain according to the power rate of received signal.Particularly, automatic gain control circuit is that low noise amplifier 2 is provided with big gain when commencing signal detects, and uses first to fourth short preamble t ₁To t ₄Determine expected gain.At beginning the 5th short preamble t ₅The time gain switched to lower gain.The gain switch level is about 20dB.

Suppose to consider the influence of multichannel and do not use the first and second short preamble t ₁And t ₂Be used for Frequency offset estimation.In this case, for example, as shown in figure 26, expression changes before the gain of low noise amplifier 2 remarkable each other different with the amplitude of the Frequency offset estimation vector of frequency shift (FS) afterwards (from multiplier 305 outputs) in the complex space.That is, about short preamble symbol t ₃And t ₄The absolute value of output of multiplier 305 bigger, and about short preamble symbol sign indicating number t ₅And t ₁₀The absolute value of output of multiplier 305 less.About preamble t ₃And t ₄The output of multiplier 305 only be 15 samplings.Therefore, residual being included in memory element 310 about preamble t with bigger value of value than first ₃And t ₄The output of multiplier 305 in equation (13) in second and the 3rd composite vector.

In structure shown in Figure 21, all short preamble t ₃To t ₁₀The summation of cross correlation results be used to Frequency offset estimation.It is big more to be used to calculate autocorrelative hits, and second and the 3rd the distortion that depends in the equation (13) is more little.Yet, if the gain of low noise amplifier 2 is changed, because about short preamble symbol sign indicating number t ₃And t ₄Sample amplitudes obviously greater than the amplitude of other preamble symbol, so in estimate vector the residual distortion vector that depends on second and the 3rd.As a result, by will be corresponding to short preamble symbol t ₅And t ₁₀79 sampling orders of output of multiplier 205 and the output addition of the memory element 207 and estimation composite vector (referring to Figure 27) that obtains causes the deterioration of Frequency offset estimation precision.

On the contrary, in the structure of frequency offset estimator shown in Figure 25 6, when receiving corresponding preamble symbol, to be weighted each preamble symbol estimated frequency shift according to the gain that in low noise amplifier 2, is provided with, and the frequency shift (FS) of weighting will be obtained final frequency offseting value mutually.Therefore, after the gain that changes low noise amplifier 2 by the bigger factor, (perhaps before changing, the gain of Frequency offset estimation do not use sampling), can be by sampling being weighted the IQ unbalanced component that reduces relatively to be included in the estimated frequency deviant, and can finally obtain more accurate frequency shift (FS).

Particularly, the absolute value of the output signal of multiplier 305 is inputed to coefficient calculation circuit 307, and coefficient of correspondence is inputed to multiplier 306.Coefficient calculation circuit 307 uses in first kind of the following stated or the second method any to calculate weighted factor corresponding to the absolute value of the output signal of multiplier 305.Basically be equal to calculating corresponding to the calculating of the weighted factor of the absolute value of the output signal of multiplier 305 corresponding to the weighted factor of the gain amplitude that in low noise amplifier 2, is provided with.

Below first method will be described.

Determine threshold value.If the absolute value of the output signal of multiplier 305 surpasses threshold value, then coefficient is set to 0.If absolute value is no more than threshold value, then coefficient is set to 1.For example, from received signal intensity indication (RSSI), determine threshold value.In this case, be offset according to following equation (14) estimated frequency:

$\mathrm{\&Delta;f}=\arg \left\{\underset{i=5N/4+1}{\overset{\mathrm{ION}}{\mathrm{\&Sigma;}}}\hat{d}(i+\frac{N}{4}){\hat{d}}^{*}\left(i\right)\right\}/(2\mathrm{\&pi;N}/4)\&CenterDot;\&CenterDot;\&CenterDot;\left(14\right)$

In other words, in first method, after the gain of low noise amplifier 2 changes from lacking preamble t ₅To t ₁₀Middle estimated frequency skew.In this case, be used to calculate autocorrelative hits and increase, and can reduce to depend on the unbalanced distortion item number of IQ.

Below second method will be described.

The inverse of the absolute value of the output signal of multiplier 305 (reciprocal) is used as coefficient.In this case, according to following equation estimated frequency skew:

$\mathrm{\&Delta;f}=\arg \{\underset{i=2N/4+1}{\overset{3N}{\mathrm{\&Sigma;}}}\frac{\hat{d}(i+\frac{N}{4}){\hat{d}}^{*}\left(i\right)}{\left|\hat{d}(i+\frac{N}{4}){\hat{d}}^{*}\left(i\right)\right|}+\underset{j=5N/4+1}{\overset{10N}{\mathrm{\&Sigma;}}}\hat{d}(j+\frac{N}{4}){\hat{d}}^{*}\left(j\right)\}/(2\mathrm{\&pi;N}/4)\&CenterDot;\&CenterDot;\&CenterDot;\left(15\right)$

As shown in figure 26, suppose that herein the gain of low noise amplifier 2 is at short preamble t ₄And t ₅Between instantaneous change.By this processing, can reduce because the IQ imbalance appears at the influence of the distortion component in adder 302 outputs.

Figure 28 shows the estimation composite vector of using first method to obtain by frequency offset estimator 6.As shown in figure 28, deleted the short preamble symbol t of use before the gain of low noise amplifier 2 reduces ₃And t ₄The estimate vector that obtains.Therefore, different with example shown in Figure 27, with the estimate vector addition of enough numbers sampling, thereby increase in the equation (13) first, simultaneously with respect to first distortion component that reduces in second to the 4th.

Figure 29 shows the Frequency offset estimation exact value (mean square error is to standardized frequency offseting value) of the use first method acquisition of comparing with the Frequency offset estimation exact value that uses structure shown in Figure 21 to obtain.

In receiver shown in Figure 1, to dispose difference filter 5 and frequency offset estimator 6 with mode shown in Figure 25.Therefore, exist IQ uneven and the time become under the situation that DC is offset, can realize accurate Frequency offset estimation by simple signal processing.

Second embodiment

Even according to the direct-conversion OFDM receiving system of first embodiment use exist IQ uneven and the time become under the situation of DC skew also the technology that can be offset by the accurate estimated frequency of simple signal processing.Be the Frequency offset estimation that provides suitable according to the device purpose of first embodiment, but and be to carry out the IQ disequilibrium regulating together with frequency offset correction.

Yet IQ disequilibrium regulating and simple frequency offset correction all fall within the scope of the invention basically.Below will be described in and be used to proofread and correct the unbalanced OFDM receiving system of IQ without departing from the present invention.

Figure 33 and Figure 34 show the structure of OFDM receiving system that is used to realize above-mentioned functions according to second embodiment of the invention.Figure 33 and Figure 34 correspond respectively to Fig. 1 and Figure 21, and are represented by identical reference number with Fig. 1 and similar component shown in Figure 21.

As Figure 33 and shown in Figure 34, be configured to be input to frequency offset estimator 6 and the uneven estimator 1000 of IQ from the signal of difference filter 5 outputs according to the OFDM receiving system of second embodiment.

Pass through above equation (12) expression from the output signal of difference filter 5, and determine by following equation with respect to the signal of the individual sampling of output signal delay N/4 (=16):

$d(i+\frac{N}{4})=\mathrm{\&phi;}(r(i+\frac{N}{4})-r(i+\frac{N}{4}-1))+{\mathrm{\&psi;}}^{*}{(r(i+\frac{N}{4})-r(i+\frac{N}{4}-1))}^{*}$

$=\mathrm{\&phi;}(r\left(i\right)-r(i-1))\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\frac{N}{4}\right)+{\mathrm{\&psi;}}^{*}{(r\left(i\right)-r(i-1))}^{*}\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\frac{N}{4})$

$=\mathrm{\&phi;\&eta;\&gamma;}+{\mathrm{\&psi;}}^{*}{\mathrm{\&eta;}}^{*}{\mathrm{\&gamma;}}^{-1}\&CenterDot;\&CenterDot;\&CenterDot;\left(16\right)$

Wherein, η=(r (i)-r (i-1)), and γ=exp (j2 π Δ f (N/4).

The signal that shifts to an earlier date N/4 sampling with respect to the signal of being determined by equation (12) is represented by following equation:

$d(i-\frac{N}{4})=\mathrm{\&phi;}(r(i-\frac{N}{4})-r(i-\frac{N}{4}-1))+{\mathrm{\&psi;}}^{*}{(r(i-\frac{N}{4})-r(i-\frac{N}{4}-1))}^{*}$

$=\mathrm{\&phi;}(r\left(i\right)-r(i-1))\exp (-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\frac{N}{4})+{\mathrm{\&psi;}}^{*}{(r\left(i\right)-r(i-1))}^{*}\exp \left(\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20071010305900691.png,A20071010305900721.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;\&Delta;f}\frac{N}{4}\right)$

$=\mathrm{\&phi;\&eta;}{\mathrm{\&gamma;}}^{-1}+{\mathrm{\&psi;}}^{*}{\mathrm{\&eta;}}^{*}\mathrm{\&gamma;}\&CenterDot;\&CenterDot;\&CenterDot;\left(17\right)$

In equation (16) and (17), the value of obtaining γ from above equation (13), and determine by the operation of the phase detecting circuit in the frequency offset estimator 6 208.To feed back to the uneven estimator 1000 of IQ by the value branches that phase detecting circuit 208 is determined, with unknown number of minimizing from equation (16) and (17), and equation (16) and (17) can be used as the function of value d, φ and η.As a result, should be appreciated that, obtain three equatioies (that is, equation (12), (16) and (17)) with respect to three known variables (φ, η and d), and can obtain separating of all variablees.

Three samplings corresponding to each equation (12), (16) and (17) stand by following equation (18) and (19) given operation:

$\frac{d(i-\frac{N}{4})-d\left(i\right){\mathrm{\&gamma;}}^{-1}}{\mathrm{\&gamma;}-{\mathrm{\&gamma;}}^{-1}}={\mathrm{\&psi;}}^{*}{\mathrm{\&eta;}}^{*}\&CenterDot;\&CenterDot;\&CenterDot;\left(18\right)$

$\frac{d(i+\frac{N}{4})-d\left(i\right)\mathrm{\&gamma;}}{\mathrm{\&gamma;}-{\mathrm{\&gamma;}}^{-1}}=\mathrm{\&phi;\&eta;}\&CenterDot;\&CenterDot;\&CenterDot;\left(19\right)$

Therefore, can from equation (18) and (19), obtain relation by following equation (20) expression:

$\frac{{\mathrm{\&psi;}}^{*}}{{\mathrm{\&phi;}}^{*}}=\frac{d(i-\frac{N}{4})-d\left(i\right){\mathrm{\&gamma;}}^{-1}}{{(d\left(i\right)\mathrm{\&gamma;}-d(i+\frac{N}{4}))}^{*}}=\mathrm{\&epsiv;}\&CenterDot;\&CenterDot;\&CenterDot;\left(20\right)$

Owing to from above equation (11), represent to be worth φ and ψ by following equation:

$\mathrm{\&phi;}=\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{j\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}$

$\mathrm{\&psi;}=\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}+j\sin \frac{\mathrm{\&theta;}}{2}$

Therefore, by approximate above equation, value φ and η are represented by following equation:

$\mathrm{\&phi;}=\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{j\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}\&ap;1+\mathrm{j\&alpha;}\frac{\mathrm{\&theta;}}{2}$

$\mathrm{\&psi;}=\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}+j\sin \frac{\mathrm{\&theta;}}{2}\&ap;\mathrm{\&alpha;}+j\frac{\mathrm{\&theta;}}{2}$

As mentioned above, in above equation, α and θ represent:

(i) amplitude of α=I component and Q component; And

The (ii) phase difference between θ=cosine signal and the sinusoidal signal.

When the I component of signal and Q component have following two kinds when concerning, the IQ imbalance appears:

(a) produce through 90 ° of phase shifters and input to the local signal (that is the output signal of local oscillator 11) that frequency mixer 3 is used for frequency inverted by being divided into two signals and signal from the output signal of phase-locked loop (PLL).If the output signal from PLL is a high-frequency signal, then phase shifter also not exclusively has 90 ° phase shift (that is, the I component of signal and Q component are non-orthogonal each other), causes occurring phase difference θ.

(b), between the I component of the input of A/D converter and Q component, difference of vibration occurs, and value α is not 0 because by the caused loss of phase shifter, about the gain error between the amplifier of I and Q component etc.

According to these relations, based on three samplings d (i-N/4), d (i) and d (i+N/4) determined value α and θ, and proofread and correct the complex baseband signal that receives, thereby proofread and correct the IQ imbalance based on value α that determines and θ.In OFDM receiving system according to second embodiment, uneven estimator 1000 determined value α of IQ and θ are (promptly, the uneven estimator 1000 estimating I/Q imbalances of IQ), and IQ disequilibrium regulating device 1100 multiply by correction coefficient corresponding to determined value with the complex baseband signal that is received, to carry out the IQ disequilibrium regulating.

The concrete technology of using uneven estimator 1000 determined value α of IQ and θ below will be described.The suitable expression formula of φ and ψ is brought in the equation (20), generates following equation:

$\mathrm{\&epsiv;}=({\mathrm{\&epsiv;}}_I+j{\mathrm{\&epsiv;}}_Q)=\frac{{(\mathrm{\&alpha;}+j\frac{\mathrm{\&theta;}}{2})}^{*}}{{(1+\mathrm{j\&alpha;}\frac{\mathrm{\&theta;}}{2})}^{*}}=\frac{(\mathrm{\&alpha;}-j\frac{\mathrm{\&theta;}}{2})}{(1-\mathrm{j\&alpha;}\frac{\mathrm{\&theta;}}{2})}\&CenterDot;\&CenterDot;\&CenterDot;\left(21\right)$

Be real part and imaginary part solve equation (21), generate following equation:

$\mathrm{\&alpha;}={\mathrm{\&epsiv;}}_I+{\mathrm{\&epsiv;}}_Q\mathrm{\&alpha;}\frac{\mathrm{\&theta;}}{2}\&CenterDot;\&CenterDot;\&CenterDot;\left(22\right)$

$\frac{\mathrm{\&theta;}}{2}=-{\mathrm{\&epsiv;}}_Q+{\mathrm{\&epsiv;}}_I\mathrm{\&alpha;}\frac{\mathrm{\&theta;}}{2}\&CenterDot;\&CenterDot;\&CenterDot;\left(23\right)$

According to equation (22) and (23), determined value α and θ satisfy following relation:

${\mathrm{\&epsiv;}}_I{\mathrm{\&alpha;}}^2+(-{\mathrm{\&epsiv;}}_I^2-{\mathrm{\&epsiv;}}_Q^2-1)\mathrm{\&alpha;}+{\mathrm{\&epsiv;}}_1=0$

$\mathrm{\&alpha;}=\frac{-(-{\mathrm{\&epsiv;}}_I^2-{\mathrm{\&epsiv;}}_Q^2-1)-\sqrt{{(-{\mathrm{\&epsiv;}}_I^2-{\mathrm{\&epsiv;}}_Q^2-1)}^2-4{\mathrm{\&epsiv;}}_I^2}}{2{\mathrm{\&epsiv;}}_I}\&CenterDot;\&CenterDot;\&CenterDot;\left(24\right)$

$\mathrm{\&theta;}=2\left(\frac{-{\mathrm{\&epsiv;}}_Q}{1-{\mathrm{\&epsiv;}}_I\mathrm{\&alpha;}}\right)\&CenterDot;\&CenterDot;\&CenterDot;\left(25\right)$

Wherein, ε determines by equation (20), and obtains variable d in the equation (20) as the output signal from difference filter 5.Because value γ determines by phase detecting circuit 208, so should be understood that determined value ε in the output signal (that is, value d) of difference filter 5 always, has therefore determined signal (that is value γ) and value α and θ from phase detecting circuit 208 feedbacks.

In a second embodiment, the uneven estimator 1000 of IQ is carried out aforesaid operations with determined value α and θ, and will export IQ disequilibrium regulating device 1100 corresponding to the coefficient correlation of determined value to.As a result, use IQ disequilibrium regulating device that the complex baseband signal that receives be multiply by correction coefficient, proofreaied and correct the IQ imbalance that in the complex baseband signal that receives, causes.

IQ disequilibrium regulating device 1100 can be set at any position between ADC 4 and the DFT 8.The IQ disequilibrium regulating device 1100 that empirical discovery is arranged on frequency offset corrector 7 upstreams provides the receiving feature that improves.In addition, experience points out, when carrying out the IQ disequilibrium regulating towards difference filter 5 with respect to the upstream of breakout x and correction signal given difference filter 5, compares with the situation of carrying out the IQ disequilibrium regulating with respect to the downstream of breakout x and to have improved receiving feature.Given this, though being configured to IQ disequilibrium regulating device 1100 according to the OFDM receiving system of second embodiment is not arranged between ADC 4 and the breakout x (referring to Figure 33), if but desired advantages still can be realized in the position of change IQ disequilibrium regulating device 1100.Therefore, the present invention is implemented in the position that can not consider IQ disequilibrium regulating device 1100.

Can use any method to determine correction coefficient based on value α and θ.For example, the uneven estimator 1000 of IQ can be determined the correction coefficient corresponding to determined value α and θ experimentally, and the table with corresponding relation between experiment value and value α and the θ can be stored in the uneven estimator 1000 of IQ.Alternatively, can from value α and θ, directly determine correction coefficient.In this case, can use following method.That is,, determine the complex baseband signal of reception by following equation according to above equation (11):

$\hat{r}\left(i\right)={\hat{r}}_I\left(i\right)+j{\hat{r}}_Q\left(i\right)$

$=(\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{j\&alpha;}\sin \frac{\mathrm{\&theta;}}{2})(r_I\left(i\right)+j{r}_Q\left(i\right))+(\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}-j\sin \frac{\mathrm{\&theta;}}{2})(r_I\left(i\right)-j{r}_Q\left(i\right))$

$=(\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2})r_I\left(i\right)+(-\sin \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2})r_Q\left(i\right)$

$+j(-\sin \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2})r_I\left(i\right)+j(\cos \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2})r_Q\left(i\right)$

Therefore, determined to satisfy the correction coefficient of following formula, and the complex baseband signal that receives multiply by definite correction coefficient to proofread and correct the IQ imbalance:

$\left(\begin{array}{c}{\hat{r}}_I\left(i\right)\\ {\hat{r}}_Q\left(i\right)\end{array}\right)=\left(\begin{array}{cc}\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}& -\sin \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}\\ -\sin \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}& \cos \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}\end{array}\right)\left(\begin{array}{c}{r}_I\left(i\right)\\ r_Q\left(i\right)\end{array}\right)$

$\left(\begin{array}{c}{r}_I\left(i\right)\\ r_Q\left(i\right)\end{array}\right)={\left(\begin{array}{cc}\cos \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}& -\sin \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}\\ -\sin \frac{\mathrm{\&theta;}}{2}+\mathrm{\&alpha;}\sin \frac{\mathrm{\&theta;}}{2}& \cos \frac{\mathrm{\&theta;}}{2}-\mathrm{\&alpha;}\cos \frac{\mathrm{\&theta;}}{2}\end{array}\right)}^{-1}\left(\begin{array}{c}{\hat{r}}_I\left(i\right)\\ {\hat{r}}_Q\left(i\right)\end{array}\right)\&CenterDot;\&CenterDot;\&CenterDot;\left(26\right)$

The structure of other parts and operation are similar with first embodiment.

Therefore, in OFDM receiving system according to second embodiment, the uneven estimator 1000 of IQ is determined correction coefficient based on the reception complex baseband signal from difference filter 5, and IQ disequilibrium regulating device 1100 multiply by definite corrected value with the complex baseband signal that receives, and is included in the IQ imbalance in the complex baseband signal of reception with correction.

By making in this way, the OFDM receiving system has been realized Figure 35 and expectation MSE characteristic shown in Figure 36.Experiment value shown in Figure 35 and Figure 36 is illustrated under the situation that does not change LAN 2 gain the MSE that obtains when α=0.05 and θ=5 °.MSE during Figure 35 value of showing α estimates, and the MSE in Figure 36 value of showing θ estimation.

Though described the IQ disequilibrium regulating method of structure shown in Fig. 1 of being used for first embodiment and Figure 21, can realize the IQ disequilibrium regulating with other device configuration equally.Other example of the structure of difference filter 5 and frequency offset estimator 6 below will be described.

Figure 37 shows another example of difference filter 5 and frequency offset estimator 6 structures.In structure shown in Figure 37, the uneven estimator 1000 of IQ is added in the structure of difference filter shown in Figure 25 5 and frequency offset estimator 6.In Figure 37, represent by identical reference number with similar parts shown in Figure 25.

As shown in figure 37, according to second embodiment, structure as shown in figure 34, will be (promptly from difference filter 5, delay cell 301 and adder 302) output signal input to the uneven estimator 1000 of IQ, and will give IQ uneven estimator 1000 corresponding to the signal feedback of value γ from phase detecting circuit 311.The uneven estimator 1000 of IQ is carried out the operation that is provided to (25) by equation (16) based on input signal, and definite disequilibrium regulating coefficient.IQ disequilibrium regulating device 1100 multiply by definite correction coefficient with the complex baseband signal that is received, to proofread and correct caused IQ imbalance in the complex baseband signal that receives.The operation of other parts and shown in Figure 25 similar, thereby omit its detailed description.

The modification of second embodiment

First revises

To shown in Figure 41, can carry out various modifications as Figure 38 to OFDM receiving system shown in Figure 33.Figure 38 shows the OFDM receiving system that is configured to the structure that is used for the IQ disequilibrium regulating is added to OFDM receiving system shown in Figure 4.Figure 39, Figure 40 and Figure 41 show the OFDM receiving system that is configured to the structure that is used for the IQ disequilibrium regulating is added to respectively Fig. 5, Fig. 6 and OFDM receiving system shown in Figure 7.

At Figure 38 to arbitrary structure shown in Figure 41, (i) will input to the uneven estimator 1000 of IQ from the output signal (referring to Figure 34) of difference filter 5, and (ii) will feed back to the uneven estimator 1000 of IQ from phase detecting circuit 208 corresponding to the signal of value γ.The uneven estimator 1000 of IQ is carried out by the given operation of equation (16) to (25) to determine correction coefficient based on input signal.IQ disequilibrium regulating device 1100 is proofreaied and correct caused IQ imbalance in the complex baseband signal that receives based on correction coefficient.Owing to add the structure that is used for the IQ disequilibrium regulating to OFDM receiving system shown in Figure 4, other structure of the OFDM receiving system shown in Figure 38, Figure 39, Figure 40 and Figure 41 and operation are similar with the OFDM receiving system shown in Fig. 4, Fig. 5, Fig. 6 and Fig. 7 respectively.

As mentioned above, to arbitrary structure shown in Figure 41, IQ disequilibrium regulating device 1100 can be arranged on any position between ADC 4 and the DFT 8 at Figure 38.Yet, to structure shown in Figure 41, use following method to improve receiving feature at Figure 38.

In Figure 38 and OFDM receiving system shown in Figure 41, IQ disequilibrium regulating device 1100 is arranged on ADC 4 and between the breakout of difference filter 5, makes and the signal through the IQ disequilibrium regulating can be inputed to difference filter 5.

In Figure 39 and OFDM receiving system shown in Figure 40,, make from the signal of DC offset corrector 10 outputs and after carrying out frequency offset correction, carry out the IQ disequilibrium regulating for following two reasons.

First reason is that the situation when carrying out the IQ disequilibrium regulating before carrying out the DC offset correction is compared, and when carrying out the IQ disequilibrium regulating after carrying out the DC offset correction, has improved receiving feature.

Second reason is to compare with the situation of carrying out frequency offset correction before carrying out the IQ disequilibrium regulating, when carrying out frequency offset correction after carrying out the IQ disequilibrium regulating, improved receiving feature.

Thereby Fig. 8 and synchronous circuit shown in Figure 9 also can be made amendment and be added the structure that is used for the IQ disequilibrium regulating.Figure 42 and Figure 43 show the circuit structure that is configured to the structure that is used for the IQ disequilibrium regulating is added to respectively the synchronous circuit of Fig. 8 and synchronous circuit shown in Figure 9.In Figure 42 and Figure 43, carry out the parts of identical function and operation with Fig. 8 and parts shown in Figure 9 and represent by the reference number identical with reference number shown in Figure 9 with Fig. 8.Below the synchronous circuit shown in Figure 42 and Figure 43 will be described.

In synchronous circuit shown in Figure 42, IQ disequilibrium regulating device 1100 is arranged between DC offset corrector and the frequency offset corrector 24.Output signal from HPF 21 is input to frequency offset estimator 22 and bag detector and thick timing detector 23, and inputs to the uneven estimator 1000 of IQ.With the OFDM receiving system shown in Figure 33, be input to the uneven estimator 1000 of IQ from the signal corresponding to value γ of frequency offset estimator 22, the uneven estimator 1000 of IQ is carried out by the given operation of equation (16) to (25) to determine correction coefficient based on input signal then.The correction coefficient of determining is offered IQ disequilibrium regulating device 1100 from the uneven estimator 1000 of IQ, the complex baseband signal that receives is carried out the IQ disequilibrium regulating based on correction coefficient.

As mentioned above, comprise the guiding high pass filter 21 that is used for each I axle and Q axle input signal and the path of DC offset estimator 25 corresponding to the synchronous circuit shown in Figure 8 of Figure 42, and open or close two switches 26 and 27 specially between two paths, to switch.Under the situation in the path of the guiding HPF 21 in activating these paths, the uneven estimator 1000 of IQ is determined correction coefficient.Other structure (for example, switch 26 and 27 switching controls) is similar to synchronous circuit shown in Figure 8, thereby omits its detailed description.

Below synchronous circuit shown in Figure 43 will be described.In synchronous circuit shown in Figure 43, switch 26 and 27 does not directly switch in response to the detection signal of bag detector and thick timing detector 23, but on-off controller 28 is set in addition.In this structure, IQ disequilibrium regulating device 1100 is arranged between DC offset corrector and the frequency offset corrector 24 equally, and the complex baseband signal that receives is carried out the IQ disequilibrium regulating.

As Figure 42 and shown in Figure 43, replace and only use STS to carry out DC offset correction, IQ disequilibrium regulating and frequency offset correction, can realize using LTS to realize the function of more accurate frequency offset correction.Figure 44 and Figure 45 show the circuit structure of the peripheral synchronous circuit that is used to carry out this function.Figure 44 and Figure 45 illustrate the diagrammatic sketch that is configured to the structure that is used for the IQ disequilibrium regulating is added to the synchronous circuit structure of Figure 13 and synchronous circuit shown in Figure 14, and are represented by Figure 13 and identical reference number shown in Figure 14 with the parts of Figure 13 and parts shown in Figure 14 execution identical function and operation.

In synchronous circuit shown in Figure 44, be provided for uneven estimator 1400 of IQ and the IQ disequilibrium regulating device 1500 of LTS independently with uneven estimator 1200 of the IQ that is used for STS and IQ disequilibrium regulating device 1300.The uneven estimator 1200 of IQ that is used for STS is identified for using short preamble to carry out the correction coefficient of thick IQ disequilibrium regulating.The uneven estimator 1400 of IQ that is used for LTS is identified for using long preamble T ₁And T ₂Determine to carry out the correction coefficient of thin IQ disequilibrium regulating.

As mentioned above, in the end of short preamble period, the IQ input is switched to the path of guiding DC offset corrector from the path of guiding high pass filter 21.After LTS, IQ disequilibrium regulating device 1300 multiply by the correction coefficient of determining by the complex baseband signal that will be received before the STS end carries out the IQ disequilibrium regulating.After this, IQ disequilibrium regulating device 1500 uses LTS to proofread and correct the IQ imbalance of less important (residual).

In circuit structure shown in Figure 44, the circuit module of the part of the reception complex baseband signal after the LTS being carried out the IQ disequilibrium regulating is set separately.Alternatively, as shown in figure 45, uneven estimator 1600 of IQ and IQ disequilibrium regulating device 1700 can use STS to carry out the IQ disequilibrium regulating, but also can use LTS to carry out the IQ disequilibrium regulating.

Describe the present invention in detail with reference to specific embodiments of the invention.Yet, should be appreciated that in the case without departing from the scope of the present invention, those skilled in the art can make various changes and modifications embodiment.

Though the embodiment of Miao Shuing is under the background of the wireless communication system that meets IEEE 802.11 standards herein, scope of the present invention is not limited thereto.Can also be used in according to the receiver of the embodiment of the invention and to repeat in the preamble portion to transmit in the wireless communication system of identical OFDM symbol, wherein, the DC subcarrier is set to null symbol to realize accurate Frequency offset estimation.Be not only the application of WLAN, and also fall within the scope of the invention such as the various digital communication technologies based on the OFDM transmission mode of received terrestrial digital broadcasting system, the 4th third-generation mobile communication system and power-line carrier communication system.

Though can overcome caused DC offset problem in the direct conversion receiver by embodiments of the invention, scope of the present invention is not limited thereto.The receiver that uses other frequency conversion method that the RF received signal is carried out down-conversion can be used for handling DC skew and frequency offset issues.

Should understand, the present invention who has disclosed with the form of exemplary embodiment, and the disclosure content in this specification is not to be used to limit the scope of the invention.True scope of the present invention should be determined according to claims.

It should be appreciated by those skilled in the art, multiple modification, combination, recombinant and improvement to be arranged, all should be included within the scope of claim of the present invention or equivalent according to designing requirement and other factors.

Claims (31)

1. a radio communication device is used to receive the bag that is made of the signal by the OFDM modulation, comprising:

Band pass filter, the orthogonal frequency-division multiplex singal of extraction desired frequency band;

Low noise amplifier has the gain of controlling according to the intensity of received signal, to amplify the described orthogonal frequency-division multiplex singal of described desired frequency band;

Frequency converter down-converts to baseband signal with the orthogonal frequency-division multiplex singal that is amplified;

Analog to digital converter converts described baseband signal to digital signal;

First high pass filter is removed the DC skew from the predetermined preamble described baseband signal partly corresponding to described bag;

Frequency offset estimator is therefrom removed estimated frequency skew the sampled signal of described baseband signal of described DC skew from forming by described first high pass filter;

Frequency offset corrector is removed estimated frequency shift (FS) from described baseband signal; And

Demodulator, demodulation is arranged in the subcarrier signal in the frequency domain from the described baseband signal that has compensated described frequency shift (FS).

2. radio communication device according to claim 1, wherein, described first high pass filter comprises differential amplifier.

3. radio communication device according to claim 1, wherein, in case detect the quick change of DC skew, described first high pass filter is just to described frequency offset estimator input detection signal; And

Described frequency offset estimator is not carried out Frequency offset estimation to the sampled signal that obtains from described first high pass filter in the described sampled signal when importing described detection signal.

4. radio communication device according to claim 1, wherein, the described orthogonal frequency-division multiplex singal that inputs to described radio communication device does not comprise the DC subcarrier; And

Described frequency offset estimator uses the preamble of two OFDM symbols of transmission to estimate skew.

5. radio communication device according to claim 4, wherein, an OFDM symbol is made of n subcarrier,

When i sampling of the time waveform of the OFDM symbol of described two transmission represented by s (i), the sampling of the OFDM symbol of first transmission is by { s (0), s (1), ..., s (n-1) } expression, the sampling of the OFDM symbol of second transmission is by { s (n), s (n+1) ..., s (2n-1) } expression, described frequency shift (FS) is represented by Δ f, and described DC skew provides described baseband signal by following equation (1) when being represented by D, and described first high pass filter is carried out by the given operation of following equation (2) described baseband signal, and output sampled signal d (i), and

Described frequency offset corrector uses described sampled signal d (i) execution to estimate described frequency shift (FS) Δ f by the given operation of following equation (3):

r(i)＝s(i)exp(j2πΔfi)+D …(1)

d(i)＝r(i+1)-r(i)

＝s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi) …(2)

$d(i+n)/d\left(i\right)=\frac{s(i+1+n)\exp \left(j2\mathrm{\&pi;f}(i+1+n)\right)-s(i+n)\exp \left(j2\mathrm{\&pi;f}(i+n)\right)}{s(i+1)\exp \left(j2\mathrm{\&pi;f}(i+1)\right)-s\left(i\right)\exp \left(j2\mathrm{\&pi;f}\left(i\right)\right)}$

$=\exp \left(j2\mathrm{\&pi;\&Delta;f}\left(n\right)\right)---\left(3\right).$

6. radio communication device according to claim 5, wherein, the DC skew change when the absolute value of the sampled signal d that is exported (i) because by D (i+1)-D (i) between given i and (i+1) individual sampling and when very big, according to following equation:

d(i)＝r(i+1)-r(i)

＝{s(i+1)exp(j2πΔf(i+1))-s(i)exp(j2πΔfi)}+{D(i+1)-D(i)}

…(4)

Described first high pass filter is exported the not detection signal of estimated frequency skew by the operation that is provided by described equation (3) is carried out in described i sampling to described frequency offset estimator.

7. radio communication device according to claim 1, wherein, described frequency converter is according to direct translative mode, uses the local frequency that produced by local oscillator that the described amplification orthogonal frequency-division multiplex singal of described desired frequency band is converted to baseband signal, and

According to the phase place of the described frequency shift (FS) counter-rotating of estimating by described frequency offset estimator by the described local frequency of described local oscillator vibration.

8. radio communication device according to claim 1 further comprises:

The DC offset estimator is estimated by the DC skew in the digital baseband signal of described analog to digital converter conversion; And

The DC offset corrector is removed the DC skew of estimating from the digital baseband signal of being changed.

9. radio communication device according to claim 8, wherein, in case detect the quick change of DC skew, described first high pass filter is just to described DC offset estimator input detection signal; And

Described DC offset estimator is got rid of the estimation DC deviant of determining before the described detection signal of input, and reappraises the DC skew.

10. radio communication device according to claim 1 further comprises second high pass filter, and the described baseband signal of exporting from described frequency converter is carried out filtering,

Wherein, described analog to digital converter will convert digital signal to by the described baseband signal of described second high pass filter transmission.

11. radio communication device according to claim 10, wherein, the cut-off frequency of described second high pass filter is set to be lower than the cut-off frequency of described first high pass filter.

12. radio communication device according to claim 8 further comprises the detector that the output signal of using described first high pass filter is carried out the bag detection and slightly regularly detected.

13. radio communication device according to claim 12 further comprises the switch that the output of described analog to digital converter is connected to specially the path of the path of described first high pass filter of guiding or the described DC offset corrector that leads.

14. radio communication device according to claim 13, wherein, when described detector detected the described predetermined preamble part that is used for Frequency offset estimation terminal, described switch switched to the described output of described analog to digital converter the path of the described DC offset corrector of guiding from the path of described first high pass filter that leads.

15. radio communication device according to claim 14, wherein, described frequency offset estimator uses in a period of time before the end of described predetermined preamble part and has used described first high pass filter therefrom to remove the described baseband signal estimated frequency skew of described DC skew.

16. radio communication device according to claim 14, wherein, described DC offset estimator is estimated the DC skew in a period of time before the end of described predetermined preamble part, and

Proofread and correct estimated DC skew the part of the described baseband signal of described DC offset corrector after the end of described predetermined preamble part.

17. radio communication device according to claim 14, wherein, the estimated frequency skew in a period of time before the end of described predetermined preamble part of described frequency offset estimator, and

Proofread and correct estimated frequency shift (FS) the part of the described baseband signal of described frequency offset corrector after the end of described predetermined preamble part.

18. radio communication device according to claim 13, further comprise the output signal of determining described analog to digital converter correlation moving average and when described moving average surpasses predetermined threshold, control the on-off controller of the switching of described switch.

19. radio communication device according to claim 13, wherein, the bag that receives comprises short preamble portion that is made of the short training sequence with relatively large subcarrier interval and the long preamble portion that is made of the long training sequence with less relatively subcarrier interval, and

Described long preamble portion after the end of described short preamble portion begin the place, described switch switches to the described output of described analog to digital converter the path of the described DC offset corrector of guiding from the path of described first high pass filter that leads.

20. radio communication device according to claim 19, wherein, the estimated frequency skew in described short preamble portion of described frequency offset estimator, and described frequency offset corrector removes estimated frequency shift (FS) from described long preamble portion, and

Described DC offset estimator is estimated DC skew in described short preamble portion, and described DC offset corrector removes estimated DC and be offset from described long preamble portion,

Described radio communication device also comprises:

The second frequency offset estimator, estimated frequency skew in the part of the described baseband signal after the described long preamble portion of therefrom removing the described frequency shift (FS) described short preamble portion, estimated and described DC skew, and

The second frequency offset corrector, the described frequency shift (FS) that removal is estimated by described second frequency offset estimator the described part of the described baseband signal after described long preamble portion.

21. radio communication device according to claim 19, wherein, the estimated frequency skew in described short preamble portion of described frequency offset estimator, and described frequency offset corrector removes estimated frequency shift (FS) from described long preamble portion,

Described DC offset estimator is estimated DC skew in described short preamble portion, and described DC offset corrector removes estimated DC and be offset from described long preamble portion,

The part of therefrom removing the described baseband signal after the described long preamble portion of the described frequency shift (FS) estimated and described DC skew in described short preamble portion is fed back to described frequency offset estimator, with estimated frequency skew the described part of the described baseband signal after described long preamble portion, and

Remove estimated frequency shift (FS) the described part of the described baseband signal of described frequency offset corrector after described long preamble portion.

22. according to claim 20 or 21 described radio communication devices, further comprise channel estimator, described channel estimator is from therefrom removing the described frequency shift (FS) estimated and described DC skew and having proofreaied and correct in the baseband signal of the described frequency shift (FS) of estimating in the described part of the described baseband signal after described long preamble portion and estimated channel described short preamble portion.

23. radio communication device according to claim 1 further comprises:

Homophase and the mutually uneven estimator of quadrature estimate that from therefrom removing the described sampled signal of described DC skew by described first high pass filter homophase is uneven mutually with quadrature; And

Homophase and quadrature be the disequilibrium regulating device mutually, and it is uneven mutually with quadrature to proofread and correct described homophase from described baseband signal.

24. radio communication device according to claim 23, wherein, described homophase and quadrature disequilibrium regulating device are mutually carried out homophase and quadrature disequilibrium regulating mutually to the described baseband signal of having proofreaied and correct described DC skew.

25. radio communication device according to claim 23, wherein, described frequency offset corrector from by described homophase and quadrature mutually the disequilibrium regulating device proofreaied and correct described homophase and the mutually unbalanced described baseband signal of quadrature and removed estimated frequency shift (FS).

26. radio communication device according to claim 1, wherein, transmission is used for a plurality of preamble symbol of Frequency offset estimation, and

Described frequency offset estimator will be obtained final estimated frequency deviant mutually to each estimated frequency shift (FS) in the described preamble symbol.

27. radio communication device according to claim 26 further comprises the gain controller of the described gain of regulating described low noise amplifier,

Wherein, when receiving corresponding preamble symbol, described frequency shift (FS)

Estimator is weighted each described frequency shift (FS) of estimating for described preamble symbol according to the described gain that is provided with in described low noise amplifier, and will be obtained final frequency offseting value mutually through the frequency shift (FS) of weighting.

28. radio communication device according to claim 26, wherein, described frequency offset estimator calculates weighted factor based on the absolute value of the described frequency shift (FS) of estimating at each described preamble symbol, by described weighted factor described frequency shift (FS) is weighted, and will be obtained final frequency offseting value mutually through the frequency shift (FS) of weighting.

29. radio communication device according to claim 28, wherein, described frequency offset estimator surpasses the frequency shift (FS) of predetermined threshold and weighted factor 1 is applied to the frequency shift (FS) that absolute value is no more than described predetermined threshold the described frequency shift (FS) of estimating for described preamble symbol is weighted by weighted factor 0 being applied to absolute value, and will be obtained final frequency offseting value through the frequency shift (FS) phase of weighting.

30. radio communication device according to claim 29, wherein, described frequency offset estimator is determined described predetermined threshold based on the intensity of received signal.

31. radio communication device according to claim 28, wherein, described frequency offset estimator is applied to described frequency shift (FS) by the weighted factor that will be formed by the inverse of the described absolute value of described frequency shift (FS) the described frequency shift (FS) of estimating for described preamble symbol is weighted, and will be obtained final frequency offseting value through the frequency shift (FS) phase of weighting.

Images