CN102868420A - LTE (long-term evolution) synchronous method - Google Patents

Abstract

The invention discloses a method for PSS sequence synchronization in LTE. A received PSS sequence is r (n), the position of PSS is searched by means of direct correlation if a received signal has no frequency deviation, the root and the position of the PSS are found according to a formula (1), theta refers to the position of the PSS to be found, dmt (n) refers to a time-domain sequence of the local PSS, subscript m indicates the root of the PSS sequence, t indicates conjugation, N indicates the length of the time-domain PSS sequence, an RSSI (received signal strength indicator) is used for receiving energy for a relevant sequence, and a calculation formula is shown in the original text. When the received signal has large frequency deviation, a series of supposed frequency deviations are selected and are eliminated in the received sequence, a group of relevance is made, a group of PSS relevant results is selected as a candidate set from the group of relevance, and the candidate set is used for detecting a next SSS sequence.

Description

The method that a kind of LTE is synchronous

Technical field

The invention belongs to wireless communication technology field, particularly the synchronous method of a kind of LTE.

Background technology

The characteristics such as LTE (LTE:Long Term Evolution, Long Term Evolution) as the main flow communication technology of future generation, has the transmission rate height, and the availability of frequency spectrum is high, and receiver is simple.Wherein, LTE's generally is to adopt main synchronizing sequence (PSS:Primary Synchronization Signal synchronously, main synchronizing sequence) and assist in synchronization sequence (SSS:Secondary Synchronization Signal, secondary synchronization sequences) carries out, for TDD-LTE, PSS is positioned on the 3rd the OFDM symbol of subframe 1 and subframe 6, and SSS is positioned on last OFDM symbol of subframe 0 and subframe 5; For FDD-LTE, PSS is positioned on last OFDM symbol of time slot 0 and time slot 10, and SSS is positioned on the penult OFDM symbol of time slot 0 and time slot 10.They all are positioned on 62 subcarriers in centre of frequency domain.PSS sequence one has 3 kinds of different roots and is respectively 25,29 and 34, when residential quarter ID is 0 divided by 3 remainders, choose root and be 25 PSS sequence, when residential quarter ID is 1 divided by 3 remainders, choose root and be 29 PSS sequence, when residential quarter ID is 2 divided by 3 remainders, choose root and be 34 PSS sequence.

When frequency shift (FS) is excessive, can't detect the PSS sequence in time domain, the relation of PSS Serial relation performance and frequency deviation can be with reference to the accompanying drawings 1, can find, when frequency shift (FS) surpass+/-during 9kHz, PSS correlation peak about 3dB that descends seriously influences net synchronization capability.When frequency deviation reach+/-during 13kHz, peak value is very little, in the very large situation of noise, peak value can be flooded by noise, just can't detect the correlation peak of PSS this moment.This crystal oscillator that just requires UE employing high-performance high stability is to overcome larger frequency deviation, and this is disadvantageous for reducing the UE cost.

Summary of the invention

The invention provides the synchronous method of a kind of LTE, this can solve under large offset frequency situation and carry out the difficult problem that PSS detects.

Technical scheme of the present invention is, a kind of for the synchronous method of LTE PSS sequence, receiving the PSS sequence is r (n), if do not have frequency deviation in the reception signal, then by the directly related position of seeking PSS, namely just find root and the position of PSS according to formula (1), formula (1) is:

$(\hat{m},\widehat{\mathrm{\θ}})=\arg \underset{m,\mathrm{\θ}}{\max }\left(\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\right)---\left(1\right)$

Wherein, θ is the position that needs the PSS of searching, d _Mt(n) be the time domain sequences of local PSS, subscript m represents taking root in of PSS sequence, Expression asks conjugation, N to represent the length of time domain PSS sequence, and RSSI receives the energy that is used for relevant sequence, and its computing formula is:

$\mathrm{RSSI}=\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{\left|r(n+\mathrm{\θ})\right|}^2,$

When receiving signal and have large frequency deviation, choose a series of hypothesis frequency deviation, and in receiving sequence, eliminate this hypothesis frequency deviation, do one group relevant, and from this group is relevant, choose the relevant result of one group of PSS as Candidate Set, for next step SSS Sequence Detection.

Further, choose positive set of frequency offsets and be [f ₁, f ₂..., f _M], the negative frequency deviation collection is the opposite number of positive set of frequency offsets, adds 0 frequency deviation, then total set of frequency offsets is [f _M... ,-f ₁, 0, f ₁..., f _M],

The PSS receiving sequence of having eliminated respectively each frequency deviation in this set of frequency offsets is carried out related operation, and chooses K maximum value and put into candidate list for subsequent treatment,

Its related operation expression formula is as follows:

$\left\{\begin{array}{c}\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{M}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{M}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\end{array}\right.---\left(2\right)$

In the formula, T _sIt is the sampling period.

Further, in carrying out the related operation process, with two eliminations of supposing frequency deviations of complex multiplier calculating same absolute, namely

At first utilize related algorithm to obtain real part corresponding to this hypothesis frequency deviation value and imaginary part is made as respectively real part 2 and imaginary part 2, long-pending real void after receiving sequence and the local PSS sequence conjugate multiplication is separated into real part 1 and imaginary part 1, real part 1 multiply by the long-pending conduct long-pending 1 of real part 2, real part 1 multiply by the long-pending conduct long-pending 2 of imaginary part 2, imaginary part 1 multiply by the long-pending conduct long-pending 3 of real part 2, imaginary part 1 multiply by the long-pending conduct long-pending 4 of empty step 2, with long-pending 1 with long-pending 4 and 1 real part as a result of, difference with long-pending 1 and long-pending 4 is 2 real part as a result of, with long-pending 3 with long-pending 2 and 1 imaginary part as a result of, difference with long-pending 3 and long-pending 2 is 2 imaginary part as a result of, and the result 1 who obtains correlation is $r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{1}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right),$ And the result 2 of correlation is $r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{1}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right).$

As can be known, when frequency deviation was excessive, the PSS correlation peak was very little from accompanying drawing 1, when frequency deviation when [5kHz, 5kHz] is interior, the PSS correlated performance can not be greatly affected, thereby our method that can adopt frequency deviation hypothesis frequency deviation when relevant is controlled within [5kHz, 5kHz].We can adopt the mode of conjugation hypothesis, make the frequency deviation Symmetrical of hypothesis, just can only carry out frequency deviation with a multiplier like this in relevant and eliminate, and have so just saved hardware resource.

The present invention proposes a kind of PSS correlation technique that adopts the frequency deviation hypothesis, compare with the conventional synchronization algorithm, it can effectively resist the impact that frequency deviation causes, and the frequency of hypothesis has the characteristic of conjugation, the use that can effectively save multiplier hard-wired the time.The maximum frequency deviation that the present invention can bear is [14kHz, 14kHz].

Description of drawings

Fig. 1 is the relation of PSS Serial relation peak value and frequency deviation.

Fig. 2 carries out the relevant structure chart of frequency deviation hypothesis.

Embodiment

Because the cycle of the appearance of synchronizing symbol is 5ms, if need detect the position of PSS, the data that only need to receive 5ms get final product.If the receiving sequence of 5ms is r (n), there is not frequency deviation if receive in the signal, then we can be by the directly related position of seeking PSS, shown in formula (1)

$(\hat{m},\widehat{\mathrm{\θ}})=\arg \underset{m,\mathrm{\θ}}{\max }\left(\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\right)---\left(1\right)$

In the formula, θ is the position that needs the PSS of searching, d _Mt(n) be the time domain sequences of local PSS, subscript m represents taking root in of PSS sequence,

Expression asks conjugation, N to represent that the length of time domain PSS sequence, RSSI are to receive the energy that is used for relevant sequence, its computational methods as shown in the formula

$\mathrm{RSSI}=\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{\left|r(n+\mathrm{\θ})\right|}^2$

Just can find root and the position of PSS according to formula (1).

When having very large frequency deviation, formula (1) possibly can't correctly find root and the position of PSS, thereby we need to choose a series of hypothesis frequency deviation at this moment, and in receiving sequence, eliminate this and suppose frequency deviation, doing one group is correlated with, and from this group is relevant, choose the relevant result of one group of PSS as Candidate Set, detect for next step SSS.

If the positive set of frequency offsets of choosing is [f ₁, f ₂..., f _M], then the negative frequency deviation collection can be elected the opposite number of positive set of frequency offsets as, adds the hypothesis of 0 frequency deviation, and then total set of frequency offsets is [f _M... ,-f ₁, 0, f ₁..., f _M].The receiving sequence of having eliminated respectively each frequency deviation in this set of frequency offsets is correlated with, and chooses K maximum value and put into candidate list for subsequent treatment.Its correlated expression formula is as follows:

$\left\{\begin{array}{c}\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{M}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{M}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\end{array}\right.---\left(2\right)$

In the formula, T _sIt is the sampling period.

With M=1, f ₁=9kHz, K=32 are example, suppose that then set of frequency offsets is [f ₁=-9kHz, 0, f ₁=9kHz], thereby have this moment the relevant of frequency deviation hypothesis to be

$\left\{\begin{array}{c}\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{1}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{1}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\end{array}\right.---\left(3\right)$

After the receive data of whole 5ms done relevant shown in (3) formula, therefrom choose the correlation peak of front 32 maximums, with the position θ of peak value, the frequency values f that supposes, the root m of peak value writes in the candidate list, does follow-up processing.

In carrying out correlated process, can calculate with a multiplier elimination of two hypothesis frequency deviations of same absolute, as shown in Figure 2, at first utilize related algorithm to obtain real part corresponding to this hypothesis frequency deviation value and imaginary part is made as respectively real part 2 and empty step 2, long-pending real void after receiving sequence and the local PSS sequence conjugate multiplication is separated into real part 1 and empty step 1, then with reference to the accompanying drawings 2 carry out computing after, 1 is exactly as a result $r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{1}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right)$ The result, 2 are exactly as a result $r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{1}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right)$ The result, calculate these two correlations and only used 1 complex multiplier and 4 real multipliers, and 4 real multipliers are equivalent to 1 complex multiplier, thereby altogether only used 2 complex multipliers.

Claims (3)

1. one kind is used for the synchronous method of LTE PSS sequence, receiving the PSS sequence is r (n), does not have frequency deviation if receive in the signal, then by the directly related position of seeking PSS, namely just find root and the position of PSS according to formula (1), formula (1) is:

$(\hat{m},\widehat{\mathrm{\θ}})=\arg \underset{m,\mathrm{\θ}}{\max }\left(\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\right)---\left(1\right)$

Wherein, θ is the position that needs the PSS of searching, d _Mt(n) be the time domain sequences of local PSS, subscript m represents taking root in of PSS sequence, Expression asks conjugation, N to represent the length of time domain PSS sequence, and RSSI receives the energy that is used for relevant sequence, and its computing formula is:

$\mathrm{RSSI}=\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{\left|r(n+\mathrm{\θ})\right|}^2,$

It is characterized in that,

When receiving signal and have large frequency deviation, choose a series of hypothesis frequency deviation, and in receiving sequence, eliminate this hypothesis frequency deviation, do one group relevant, and from this group is relevant, choose the relevant result of one group of PSS as Candidate Set, for next step SSS Sequence Detection.

2. as claimed in claim 1ly it is characterized in that for the synchronous method of LTE PSS sequence,

Choose positive set of frequency offsets and be [f ₁, f ₂..., f _M], the negative frequency deviation collection is the opposite number of positive set of frequency offsets, adds 0 frequency deviation, then total set of frequency offsets is [f _M... ,-f ₁, 0, f ₁..., f _M],

The PSS receiving sequence of having eliminated respectively each frequency deviation in this set of frequency offsets is carried out related operation, and chooses K maximum value and put into candidate list for subsequent treatment,

Its related operation expression formula is as follows:

$\left\{\begin{array}{c}\frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{M}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\\ .\\ .\\ .\\ \frac{{\left|\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{M}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right)\right|}^2}{\mathrm{RSSI}}\end{array}\right.---\left(2\right)$

In the formula, T _sIt is the sampling period.

3. as claimed in claim 2ly it is characterized in that for the synchronous method of LTE PSS sequence,

In carrying out the related operation process, with two eliminations of supposing frequency deviations of complex multiplier calculating same absolute, namely

At first utilize related algorithm to obtain real part corresponding to this hypothesis frequency deviation value and imaginary part is made as respectively real part 2 and imaginary part 2, long-pending real void after receiving sequence and the local PSS sequence conjugate multiplication is separated into real part 1 and imaginary part 1, real part 1 multiply by the long-pending conduct long-pending 1 of real part 2, real part 1 multiply by the long-pending conduct long-pending 2 of imaginary part 2, imaginary part 1 multiply by the long-pending conduct long-pending 3 of real part 2, imaginary part 1 multiply by the long-pending conduct long-pending 4 of empty step 2, with long-pending 1 with long-pending 4 and 1 real part as a result of, difference with long-pending 1 and long-pending 4 is 2 real part as a result of, with long-pending 3 with long-pending 2 and 1 imaginary part as a result of, difference with long-pending 3 and long-pending 2 is 2 imaginary part as a result of, and the result 1 who obtains correlation is $r(n+\mathrm{\θ})\exp \left(j2\mathrm{\π}{f}_{1}n{T}_s\right)d_{\mathrm{mt}}^{*}\left(n\right),$ And the result 2 of correlation is $r(n+\mathrm{\θ})\exp (-j2\mathrm{\π}{f}_{1}n{T}_s)d_{\mathrm{mt}}^{*}\left(n\right).$

Images