A kind of multipath searching method of broadband CDMA system
Technical field
The present invention relates in the wireless mobile communications field multipath searching method in the Direct-Spread code division multiple access system, more specifically is how to search for multipath more accurately in broadband CDMA system (WCDMA).
Background technology
In the third generation mobile system, the frequency spectrum of wireless channel is quite wide, utilizes spread spectrum and code division multiple access, and a plurality of users can use identical channel.Utilize direct spreading sequence, information symbol is represented by the sequence of chip.In the wireless interface receiving end, chip sequence is correlated with the recovering information symbol.Therefore because mobile communication is to carry out under the electric wave environment of complexity, how overcoming the multipath fading phenomenon that radio wave propagation causes is a basic problem of mobile communication.In the Direct-Spread code division multiple access system, owing to use wide-band spread spectrum, the correlation bandwidth of radiofrequency signal bandwidth ratio decline is much bigger, thereby the component that can adopt Rake receiver to tell different paths from received signal is separated the mediation merging respectively, thereby can strengthen received signal, improve signal to noise ratio.
Multipath searching becomes a key method in the mobile communication.Existing multipath searching method mainly contains two kinds: matched filtering method and sliding correlation method.The matched filter method be by with local code word as filter coefficient, carry out filtering to received signal, by the peak of output, determine some peak phases of input signal, just the position of multipath component.Sliding correlation method is the local code word phase place of sliding, and carries out related operation with received signal, therefrom seeks the peak of correlation, just obtains multipath component.The method speed of matched filtering is very fast, can synchronously provide the Energy distribution of current each code word phase place with input signal, but needs to consume a lot of hardware resources.Sliding correlation method speed is slow, and its speed depends primarily on the correlation length of correlator.
Generally all there is multipath in wireless mobile channel in the reality, and each footpath intensity also can change.The variation of multipath intensity and the existence of noise all can significantly reduce the multi-path resolved power of search energy window.Because the influence of Rayleigh fading, it is different to tend to cause decline decline in the update cycle at different multipaths, certain bar footpath multipath in the update cycle fading ratio will cause this footpath to lose seriously the time.Often the method for using is to use noncoherent accumulation, particularly each phase place energy is carried out filtering.The most frequently used is first order IIR (infinite impulse response) filter that filter coefficient is fixed as α.But corresponding different wireless channels, the rate of change difference of decline need reach optimum efficiency with different factor alpha, for example, shown in Figure 1, when translational speed is low, decline changes slow, at this moment needs to make full use of correlation between signals with bigger factor alpha, reaches performance.And translational speed is shown in Figure 3 as Fig. 2 when higher, and decline changes than very fast, at this moment needs to deacclimatize correlation relatively poor between the signal by situation with less factor alpha, avoids bringing unnecessary noise, increases and judges by accident.Conventional searcher can't be simple and effective judgement wireless channel situation.Has only balanced various situation; Get the fixed value of a compromise, can't reach the optimal effectiveness under the various conditions, at the different cycles of multipath existence, optimum factor alpha is also different simultaneously.The present invention is exactly the problems referred to above that exist at existing code division multiple address communication system multipath searching method, propose a kind of new practical and effective method and estimate the wireless channel situation, in conjunction with the life cycle of multipath, the method for filter coefficient α is set, can improve the accuracy of Multipath searching.
Summary of the invention
The present invention seeks to utilize the relation of the speed and the Doppler frequency shift of Rayleigh fading in the wireless channel, estimate the wireless channel situation, in conjunction with the life cycle of multipath, multipath to the out of phase under the different situations is provided with different iir filter factor alpha, improve the accuracy of Multipath searching and to the new reaction speed that the footpath occurs, thereby improve the overall performance of Rake receiver.
In order to realize the foregoing invention purpose, the present invention proposes a kind of multipath searching method of broadband CDMA system, i.e. binary search, for the first time, utilize conventional method to finish initial search earlier, the result who utilizes first step search to obtain estimates the Doppler frequency shift of travelling carriage.Set iir filter factor alpha size on the out of phase according to each footpath situation and Doppler frequency shift; For the second time, utilize the iir filter factor alpha of having upgraded to carry out the search second time, just can more accurately search new multipath.Concrete grammar is:
The first step: search for the first time, adopt conventional method to carry out initial Multipath searching, coherent accumulator 401 input baseband signals are done the relevant of different delay with known scrambler and pilot tone, obtain amplitude time delay profile (ADP), the ADP of 402 pairs of inputs of non-coherent accumulator carries out noncoherent accumulation, export each time delay noncoherent accumulation energy value, IIR filtering 403 is carried out filtering with energy value, and peak value selector 404 is sought peak value and found each footpath time-delay from filtered energy value.
Second step: beginning is search for the second time, adopt searcher, at first, use frequency offset estimator 505, utilize the information of searching for out strong footpath to estimate the Doppler frequency shift of travelling carriage, concrete as Fig. 6, footpath selector 601 is selected a footpath the strongest from multipath information, send into low pass filter 602, filtering noise is tried to achieve Doppler frequency shift by zero passage detection accumulator 603 and Doppler frequency shift calculator 604 then.
The 3rd step: calculate iir filter coefficient 506, in the update cycle, set iir filter factor alpha size on the out of phase according to the life span in the Doppler frequency shift that obtains previously and original path position and footpath at current multipath.
The 4th step: upgrade each phase place iir filter factor alpha, the pilot signal of current multipath in the update cycle done the relevant of different delay with known scrambler and pilot tone, obtain the amplitude time delay profile, carry out noncoherent accumulation.
The 5th step: according to each phase place iir filter factor alpha of setting, by coherent accumulator 501, non-coherent accumulator 502, iir filter 503, and find new multipath delay by the method that peak value selector 504 is searched peak value.
The 6th the step: at regular intervals, since second the step carry out once new Multipath searching, obtain new multipath delay.
Estimate the Doppler frequency of travelling carriage owing to use multipath information results estimated in this programme, estimate wireless channel environment, and then according to the iir filter factor alpha size on the wireless channel environment that obtains and original multipath information and the set positions out of phase.By dynamic setting iir filter factor alpha size, adapt to different wireless channel environments and multipath information, make Multipath searching reach rapidity and efficient.
Description of drawings
The decline situation of footpath 10ms when Fig. 1 mobile station speed is 3 kilometers;
The decline situation of footpath 10ms when Fig. 2 mobile station speed is 50 kilometers;
The decline situation of footpath 10ms when Fig. 3 mobile station speed is 120 kilometers;
Fig. 4 routine search device structure chart;
Fig. 5 improves the structured flowchart of searcher;
The structured flowchart of Fig. 6 frequency offset estimator;
Fig. 7 Rake receiver block diagram;
Fig. 8 flow chart of the present invention.
Embodiment
Enforcement to technical scheme is described in further detail below, and the technical staff in same field can be easy to realize the present invention.
At a Direct-Spread code division multiple access system, generally be to realize synchronously with pilot signal.Pilot signal is the important component part of DPCCH (Dedicated Physical Control Channel, Dedicated Physical Control Channel), knows in advance at receiving terminal under the situation of pilot signal and the scrambler that uses and searches for multipath.
The structure of Rake receiver as shown in Figure 7, baseband signal is delivered to respectively in searcher 701 and the demodulation module 704.Searcher 701 is searched multipath from baseband signal, find the time-delay in every footpath, offers demodulation module 704 respectively.Demodulation module 704 is according to each the footpath time-delay that obtains, carry out data demodulates, obtain the DPCCH (Dedicated Physical Control Channel, Dedicated Physical Control Channel) and DPDCH (Dedicated Physical Data Channel, the Dedicated Physical Data Channel) component in each footpath.The Dedicated Physical Control Channel signal in each footpath is sent to channel estimation module 702, simultaneously Dedicated Physical Control Channel and Dedicated Physical Data Channel signal is sent into high specific merging module 703.In channel estimation module 702, obtain each channel estimation value directly according to the Dedicated Physical Control Channel signal of importing, these values are sent to high specific merge in module 703 and the searcher 701.In high specific merging module, utilize each footpath channel estimation value of input, each Dedicated Physical Data Channel and control channel is directly carried out the high specific union operation, obtain the result of Dedicated Physical Data Channel and Dedicated Physical Control Channel.It is quite important to RAKE receiver that each footpath time-delay accurately is provided, otherwise can not demodulate correct data.Because the time-varying characteristics of wireless channel need searcher to bring in constant renewal in each footpath time-delay.Searcher 701 is main bodys of the present invention.
We suppose has time-delay to be T in the multipath channel
1, T
2..., T
NThe N paths, and RAKE receiver uses searcher to record the strongest M≤N paths, M demodulator is used for this strongest M paths of demodulation.The structure of searcher 701 as shown in Figure 5, it is that scrambler and the pilot tone of 2L+1 done the relevant of different delay t that baseband signal is at first delivered in the amplitude time delay profile correlator 501 with known length, obtains amplitude time delay profile A (t).For the n footpath, as time delay t=T
nThe time, a peak value can appear in the amplitude time delay profile.But owing to multipath intensity can change along with the variation of time, and because the existence of noise might cause amplitude time delay profile peak value to change.Therefore need carry out the influence that noncoherent accumulation and phase filtering are eliminated noise and channel fading variation.Amplitude time delay profile A (t) is sent to and carries out noncoherent accumulation in the non-coherent accumulator 502, is sent in each phase filter 503 again each phase place is carried out the first order IIR filtering that coefficient is a (t), and the result is sent to the time-delay that peak value selector 504 obtains each footpath.And the coefficient a (t) of iir filter is different for different phase places.The process of its setting is as follows, at first Doppler frequency deviation estimator 505 calculates Doppler frequency shift according to the channel estimation value of input, and IIR coefficient module 506 is set the size of filter coefficient a (t) on out of phase according to the Doppler frequency shift of importing and known each path position then.
The variation speed of multipath intensity is relevant with motion velocity of mobile station, when motion velocity of mobile station faster the time, the channel fading variation is than very fast in cycle at a Multipath searching, and maximum and minimum value have all occurred, and each footpath is little in the decline mean variation of different Multipath searching in the cycle.When the movement velocity of travelling carriage is slower, a Multipath searching in the cycle channel fading change slow, might all be smaller value in the one-period, and another cycle all is bigger value, the decline mean variation of each footpath in the different search cycles be bigger.Fig. 1, Fig. 2 and Fig. 3 represent that respectively the speed of travelling carriage is 3 kilometers, the decline situation when 50 kilometers and 120 kilometers in the time of 10ms one frame.As can be seen, along with the raising of speed, in the time, the speed that decline changes is faster at frame data, and speed is high more like this, and the decline mean value that comes out the footpath is accurate more.Otherwise when speed was low especially, the decline mean value that comes out the footpath was inaccurate more.Doppler frequency shift can be determined that computing formula is as follows by the speed of travelling carriage:
Wherein, f
mIt is Doppler frequency shift; V is a motion velocity of mobile station; C is the light velocity; f
cBe carrier frequency, be 800MHz or 1.9GHz.Here provide a kind of spectrum distribution of Rayleigh fading in typical case:
By following formula as can be known, the frequency spectrum of Rayleigh fading can equal f in frequency
mThe time have a maximum.Here we just need to have investigated Doppler frequency deviation estimator 505, and its module frame chart as shown in Figure 6.The channel estimation value in each footpath is sent into footpath selector 601, selects the strongest real part or the imaginary part fade (t) of energy according to each footpath real part and imaginary part average energy value.What the selection energy was the strongest is in order to reduce noise to estimation effect.The fade (t) that elects is sent in the low pass filter 602, fade (t) is carried out low-pass filtering, removal High-frequency Interference and noise, particularly carrier wave frequency deviation wherein obtains fade ' (t) to the influence of channel estimation value.Fade ' (t) send into zero passage detection accumulator 603 after, detect the zero passage number of times in the input waveform, and the zero passage number of times carried out IIR filtering, obtain the frequency n of waveform zero passage in a frame
Zero, send into Doppler frequency shift calculator 604.Calculate Doppler frequency shift f by following formula
Doppler
T wherein
FrameBe the time span of a frame, because twice zero passage in the sinusoidal wave one-period, therefore will be divided by 2.
In fact, the real part of Rayleigh fading and imaginary part can be similar to regards with the Doppler frequency deviation to be the sine wave that fundamental frequency changes as, so we will obtain an average decline in footpath accurately, only need be averaged just passable in its m period of change.For the standard sine wave signal, only need in its one-period, be averaged just passable, but because Rayleigh fading itself is a random process, just approximate to have with the Doppler frequency deviation be the sine wave characteristics of frequency, obviously average length only one-period be far from being enough.Coefficient is that the iir filter of a can be expressed as:
y(n)=a*x(n)+(1-a)*y(n-1)
It is not have partially to estimate that its variance is as calculated
Variance is that along with diminishing of a, variance diminishes with the rule of the variation of a.The iir filter of its approximately equivalent can be expressed as:
Wherein ceil (x) expression is to rounding on the x, and establishing x (n) cycle is T
x, then the time span of filtering is exactly T
x* ceil (1/a).And the approximate cycle of Rayleigh fading is exactly 1/f
Doppler, therefore work as T
x* ceil (1/a) 〉=m* (1/f
Doppler) time, the amplitude time delay function that filtering obtains just can have been represented the decline in each footpath more accurately.Therefore
Wherein floor (x) expression rounds downwards, and we just can set suitable iir filter coefficient according to the Doppler frequency shift that obtains like this.
In conventional method, the value of a is identical for different phase places, and is owing to having phase place directly and only having the variance difference of noise phase very big, very big in the variance of footpath and adjacent position amplitude time delay function thereof at that time, and less in noise position variance.Therefore we can consider in that different a values is set at different positions, and in the noise position, because variance itself is smaller, a value can so also help occurring the quick seizure in new footpath greatly.In footpath and adjacent position thereof, because variance itself is bigger, a value is smaller, and concrete size is set according to Doppler frequency shift.When Doppler frequency shift was big, a value can be too little, but when Doppler frequency shift hour, Doppler frequency shift just need be smaller.For with footpath adjacent half time slot position on a value, can get the median that a value on a value and the noise is gone up in the footpath.For emerging footpath, begin several cycle a values keep constant originally, through its amplitude time delay function of several cycles still after on the thresholding, turn a value down according to the size of Doppler frequency deviation.
The result of use channel estimating estimates the Doppler frequency of travelling carriage in this programme, and then sets iir filter factor alpha size according to the Doppler frequency shift that obtains.By the size of dynamic setting iir filter factor alpha, adapt to different wireless channel environments and multipath information, Multipath searching reaches rapidity and efficient.