CN104506466A - Multi-carrier CDMA (code division multiple access) signal spreading code and information sequence blind estimation method - Google Patents

Abstract

The invention discloses a multi-carrier CDMA signal spreading code and information sequence blind estimation simplifying method. The m multi-carrier CDMA signal spreading code and information sequence blind estimation simplifying method comprises establishing frequency domain spreading signal model, taking a time-varying channel environment into consideration, integrating theories of sectional modeling, approximately establishing joint posterior probability distribution through sequential Monte Carlo sampling particles, and performing parameter state estimation according to sampling values and importance weights; meanwhile, reasonably designing implementing processes, and correcting original iterative computation importance weights; making full use of space-time block codes and the orthogonal properties of multi-carrier modulation to simplifying the implementation processes. The multi-carrier CDMA signal spreading code and information sequence blind estimation simplifying method well meets the requirements on rapidly extracting multi-carrier CDMA signal spreading codes with space-time block codes as well as information sequences under the time-varying channel environment, greatly improves the estimation precision of spreading parameters and has relatively low computation complexity. The multi-carrier CDMA signal spreading code and information sequence blind estimation simplifying method can be directly applied to a non-cooperative spread communication system as well as corresponding systems such as software radio.

Description

A kind of CDMA multiple carrier signal spread-spectrum code and information sequence blind estimating method

Technical field

The invention belongs to non-cooperating signal of communication treatment technology in signal transacting field, specifically refer to a kind of CDMA multiple carrier signal spread-spectrum code and the information sequence blind estimating method with Space-Time Block Coding.

Background technology

Along with the development of DS-CDMA technology, traditional spread spectrum because there is contradiction between processing gain and data transfer rate, thus limits its application in high speed data transfer.In order to better meet the demand of satellite communication, external existing satellite extensively adopts WCDMA communication system, and in spread spectrum communication, introduce multicarrier (MC) direct sequence spread spectrum skill, adopt Space-Time Block Coding (STBC) to encode, to improve the transmission rate of data and to increase the capacity of system simultaneously.Therefore, under non-cooperative communication condition, the parameter Estimation such as spreading code and information sequence is carried out to the CDMA multiple carrier signal with Space-Time Block Coding there is important military significance.

At present, in non-co-operation signal process field, be still in the starting stage to the treatment research of the CDMA multiple carrier signal with Space-Time Block Coding.In " IEEE Journal on Selected Areas in Communiations " phase magazine calendar year 2001 19 " Bayesian Monte Carlo MultiuserReceiver for Space-Time Coded Multi-carrier CDMA Systems " literary composition, Yang Zigang adopts the method for Gibbs model to carry out effective estimation to this system information sequence and channel parameter under condition for cooperation, obtains good result; But for non-cooperation, existing achievement in research is mainly carried out for CDMA multiple carrier, directly can not implant and be applied in STBC-MC-CDMA system, and these methods remain to be further improved the adaptive capacity of low signal-to-noise ratio and computation complexity.Within 2010, Bangwon Seo adopts the method for least mean-square error adaptive-filtering to process, and its convergence rate is very fast, but not strong to the adaptive capacity of signal to noise ratio.On this basis, within 2012, Tsui-Tsai Lin discloses a kind of correlation matrix method for STBC-MC-CDMA system on (IEEE Transactions on Vehicular Technology), the method well can adapt to the demand of low signal-to-noise ratio, and amount of calculation is relatively low, its weak point is: rate of convergence is comparatively slow, and robustness is poor.

This shows, existing method can't meet the needs of STBC-MC-CDMA system, the channel circumstance become when simultaneously considering, also needs a kind of more effective spreading code of research and information sequence blind estimate method for simplifying.

Summary of the invention

Technical problem to be solved by this invention is, for the deficiencies in the prior art, propose a kind of STBC-MC-CDMA signal spread-spectrum code of analyzing based on sequential Monte Carlo (SMC) and information sequence blind estimating method, consider the impact of the non-linear factors such as time-variant channel environment simultaneously, it can meet the rapid extraction requirement of spreading code and information sequence in non-cooperating spread spectrum communication preferably, substantially increase the estimated accuracy of spread spectrum parameter, and there is lower computation complexity.The present invention can directly apply to non-cooperating spread spectrum communication system, also can be used for the systems such as corresponding software radio.

For solving the problems of the technologies described above, the present invention is achieved by the following technical solutions: by setting up the signal model of frequency domain spread spectrum, the channel circumstance become during consideration, in conjunction with the thought of segmentation modeling, adopt the sampling particle of SMC to be similar to the associating Posterior probability distribution set up, and carry out parameter state estimation according to sample value and importance weight size; Reasonably design performs flow process simultaneously, revises original iterative computation importance weight step, and makes full use of the orthogonal attributes of Space-Time Block Coding and multi-carrier modulation, the implementation procedure of method for simplifying.

Described spreading code and information sequence blind estimating method are mainly used in spreading code and the information sequence blind estimate of STBC-MC-CDMA signal, and the channel circumstance become during the foundation of signal model consideration, suppose that the m frame data of a kth user are propagated through multipath channel on a jth antenna, its expression formula is

$h_{k,m,j}\left(t\right)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{\mathrm{\β}}_{k,m,j,l}\mathrm{\δ}(t-\frac{l}{\mathrm{\Δl}})---\left(1\right)$

In formula: δ () is Dirac function, L is the number of path propagated, β _{k, m, j, l}for the multiple amplitude of fading of m frame data l paths on a jth antenna of a kth user, Δ _ffor the bandwidth of multicarrier system.After discrete Fourier transform, the frequency domain response of channel on the n-th subcarrier is

$H_{k,m,j}\left(n\right)\stackrel{\mathrm{\Δ}}{=}{H}_{k,m,j}\left(n{\mathrm{\Δ}}_f\right)=\underset{l=0}{\overset{L-1}{\mathrm{\Σ}}}{\mathrm{\β}}_{k,m,j,l}\exp (-\mathrm{j\πnl}/N)=w_f^H\left(n\right)h_{k,m,j}---\left(2\right)$

In formula: $h_{k,m,j}\stackrel{\mathrm{\Δ}}{=}{[{\mathrm{\β}}_{k,m,j,0},{\mathrm{\β}}_{k,m,j,1},...,{\mathrm{\β}}_{k,m,j,L-1}]}^T$ For the time-domain response of multipath, order $h_{m,j}\stackrel{\mathrm{\Δ}}{=}{[h_{1,m,j}^T,h_{2,m,j}^T,...,h_{K,m,j}^T]}^T,$ for corresponding discrete fourier variation coefficient, namely

$w_f\left(n\right)\stackrel{\mathrm{\Δ}}{=}{[1,\exp (-j2\mathrm{\πn}/N),...,\exp (-j2\mathrm{\πn}(L-1)/N)]}^H---\left(3\right)$

Suppose that channel is non-time-varying within the STBC time of a frame, and in different M, by time the amplitude of fading that becomes be modeled as AR Model of First, namely

h _m,j＝Fh _m-1,j+w _m(4)

In formula: F=α I _kL, w _mfor zero-mean, covariance matrix is multiple Gaussian Profile.

So the STBC data of getting M frame at receiving terminal carry out processing (now comprise the time domain data of 2M frame, because in a grouping data of transmission 2 frames), can obtain the discrete Received signal strength that user adds up to K is

In formula: T is alternate function, in general system, T is the elementary matrix that (N × N) ties up, v _mfor zero-mean covariance matrix is multiple Gaussian noise.And have

$C_{k,j}\stackrel{\mathrm{\Δ}}{=}\mathrm{diag}{\left\{c_{k,j}\right\}}_{N\×N},c_{k,j}={[c_{k,j}{\left(1\right)}^T,...,c_{k,j}{\left(P\right)}^T]}^T,j=\mathrm{1,2}---\left(6\right)$

$B_{k,m,j}\stackrel{\mathrm{\Δ}}{=}\mathrm{diag}\{b_{k,m,j}\⊗I_G\}=\mathrm{diag}{\{b_{k,m,j}\left(1\right)I_G,...,b_{k,m,j}\left(P\right)I_G\}}_{N\×N},j=\mathrm{1,2}---\left(7\right)$

$W_f\stackrel{\mathrm{\Δ}}{=}{[w_f\left(0\right),w_f\left(1\right),...,w_f(N-1)]}_{N\×L}^H---\left(8\right)$

In formula: represent Kronecker product, b _{k, m, j}, j=1,2 is the symbolic vector that need send in the m frame STBC moment, c _k,jfor distributing to the spreading code vector of a jth antenna.

Described segmentation modeling thought refers to: Received signal strength is carried out segment processing according to different symbol ways and number of antennas, respectively to p=1 ..., P and j=1,2 adopt SMC method to carry out iteration sampling when getting different value, and its signal model is

$\begin{array}{c}{y}_m(p,j)=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}{X}_{k,m}(p,j)W(p,j)h_{k,m,j}+v_m(p,j)\\ m=1,...,M.P=1,...,P.j=\mathrm{1,2}\end{array}---\left(9\right)$

In formula: X _k,m(p, j) is from matrix X _k,m(G × G) of middle extraction ties up submatrix, and this submatrix is by containing element b _{k, m, j}all row and column compositions of (p); In like manner y _m(p, j) .., h _{k, m, j}and v _m(p, j) also for extracting the submatrix of composition from homography.Due to all user K, alternate function T is identical, so form above formula being write as matrix multiple can obtain

y _m(p，j)＝X _m(p，j)Ψ(p，j)h _m，j+v _m(p，j)

(10)

M=1 ..., M.p=1 ..., P.j=1, in 2 formulas: h _m,jfor KL dimensional vector, and have

$X_m(p,j)\stackrel{\mathrm{\Δ}}{=}{[X_{1,m}(p,j),...,X_{K,m}(p,j)]}_{G\×\mathrm{KG}}---\left(11\right)$

For stating conveniently, be defined as follows

b _m(p,j)＝[b _1,m,j(p),b _2,m,j(p),…,b _K,m,j(p)] (13)

B _m(p,j)＝[b ₁(p,j),b ₂(p,j)…,b _m(p,j)] (14)

c(p,j)＝[c _1,j(p) ^T,c _2,j(p) ^T,…,c _K,j(p) ^T] ^T(15)

Y _m(p,j)＝[y ₁(p,j),y ₂(p,j),…,y _m(p,j)] (16)

$H_{m,j}\stackrel{\mathrm{\Δ}}{=}{[h_{1,j}^T,h_{2,j}^T,...,h_{m,j}^T]}^T---\left(17\right)$

Goal in research of the present invention is: when p, j get different value respectively, according to SMC method, from joint posterior distribution p (B _m(p, j), c (p, j), H _m,j| Y _m(p, j)) in sequential sampling obtain particle 1≤m≤M, thus estimate corresponding spreading code and information sequence vector.

According to bayesian criterion, by as follows for the Posterior distrbutionp function decomposition of unknown parameter:

p(B _m(p，j)，c(p，j)，H _m，j|Y _m(p，j))

(18)

＝p(H _m，j|B _m(p，j)，c(p，j)，Y _m(p，j))p(B _m(p，j)，c(p，j)|Y _m(p，j))

As can be seen from the above equation, two parts can be divided into the estimation of unknown parameter and carry out, kalman filter method can be adopted H for Section 1 on the right of equal sign _m,javerage and covariance carry out iteration renewal, Section 2 is the Posterior distrbutionp probability of low-dimensional, adopts SMC method to carry out sequential sampling, makes Z _m(p, j)=(B _m(p, j), c (p, j)), z _m(p, j)=(b _m(p, j), c (p, j)).

Described simple implementation process refers to: in conjunction with the execution flow process of kalman filter method, make full use of the orthogonal attributes of Space-Time Block Coding and multi-carrier modulation, simplifies iterative step, greatly reduces the computation complexity of method, specific as follows

From formula (4) and (10), within the m frame STBC time, h _m,jobey multiple Gaussian Profile, namely

$p\left(h_{m,j}\right|Z_m^{\left(i\right)}(p,j),Y_m(p,j))~N_c(h_{m|m,j}^{\left(i\right)},P_{m|m,j}^{\left(i\right)})---\left(19\right)$

Average in formula and variance can kalman filter method be passed through, provide with shi Jinhang iteration upgrades, and its process is:

$h_{m|m-1,j}^{\left(i\right)}={\mathrm{Fh}}_{m-1|m-1,j}^{\left(i\right)}---\left(20\right)$

$P_{m|m-1,j}^{\left(i\right)}={\mathrm{FP}}_{m-1|m-1,j}^{\left(i\right)}{F}^H+{\mathrm{\Σ}}_w---\left(21\right)$

${\hat{y}}_m^{\left(i\right)}(p,j)=X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)h_{m|m-1,j}^{\left(i\right)}$

${\mathrm{\υ}}_{m,j}^{\left(i\right)}=y_m^{\left(i\right)}(p,j)-{\hat{y}}_m^{\left(i\right)}(p,j)---\left(23\right)$

$P_{\mathrm{\υ},m,j}^{\left(i\right)}={\mathrm{\Σ}}_v+\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]P_{m|m-1,j}^{\left(i\right)}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H---\left(24\right)$

$G_{m,j}^{\left(i\right)}=P_{m|m-1,j}^{\left(i\right)}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H{\left[P_{\mathrm{\υ},m,j}^{\left(i\right)}\right]}^{-1}---\left(25\right)$

$h_{m|m,j}^{\left(i\right)}=h_{m|m-1,j}^{\left(i\right)}+G_{m,j}^{\left(i\right)}{\mathrm{\υ}}_{m,j}^{\left(i\right)}---\left(26\right)$

$P_{m|m,j}^{\left(i\right)}=P_{m|m-1,j}^{\left(i\right)}-G_{m,j}^{\left(i\right)}{P}_{\mathrm{\υ},m,j}^{\left(i\right)}{\left(G_{m,j}^{\left(i\right)}\right)}^H---\left(27\right)$

In formula: for status predication value, for the covariance matrix of status predication value, for the measured value of prediction, for innovation process, for newly ceasing matrix, for kalman gain.

From the definition of formula (9) and (10)

$\begin{array}{c}{\left[X_m(p,j)\mathrm{\Ψ}(p,j)\right]}^H\left[X_m(p,j)\mathrm{\Ψ}(p,j)\right]={\mathrm{\Ψ}}^H(p,j)X_m^H(p,j)X_m(p,j)\mathrm{\Ψ}(p,j)\\ =\mathrm{KG}\·I_{\mathrm{KL}}\end{array}---\left(28\right)$

The orthogonal attributes of STBC has been used in formula and the orthogonal attributes Ψ of multi-carrier modulation ^h(p, j) Ψ (p, j)=KGI _kL.Suppose for diagonal matrix, namely

$P_{m-1|m-1,j}^{\left(i\right)}={\mathrm{\δ}}_{m-1}^{\left(i\right)}{I}_{\mathrm{KL}}---\left(29\right)$

Above formula is substituted into formula (21) can obtain also be diagonal matrix, namely

$P_{m|m-1,j}^{\left(i\right)}={\mathrm{\δ}}_{m-1}^{\left(i\right)}{\mathrm{FF}}^H+{\mathrm{\Σ}}_w=({\left|\mathrm{\α}\right|}^2{\mathrm{\δ}}_{m-1}^{\left(i\right)}+{\mathrm{\σ}}_w^2)I_{\mathrm{KL}}---\left(30\right)$

Order and according to the theorem of inverting of matrix, by formula (24) newly breath inverse of a matrix write as

$\begin{array}{c}{\left[P_{\mathrm{\υ},m,j}^{\left(i\right)}\right]}^{-1}={\mathrm{\Σ}}_v^{-1}-{\mathrm{\Σ}}_v^{-1}\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]\\ \×{\left({\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H{\mathrm{\Σ}}_v^{-1}\right[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)]+{\left(P_{m|m-1,j}^{\left(i\right)}\right)}^{-1})}^{-1}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H{\mathrm{\Σ}}_v^{-1}\end{array}---\left(31\right)$

Substitution formula (28) and (30) can obtain

$\begin{array}{c}{\left[P_{\mathrm{\υ},m,j}^{\left(i\right)}\right]}^{-1}=\frac{1}{{\mathrm{\σ}}_v^2}{I}_G-\frac{1}{{\mathrm{\σ}}_v^4}\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]\\ \×{\left(\right[\frac{\mathrm{KG}}{{\mathrm{\σ}}_v^2}+\frac{1}{{\mathrm{\β}}_m^{\left(i\right)}}\left]I_{\mathrm{KL}}\right)}^{-1}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H\\ =\frac{1}{{\mathrm{\σ}}_v^2}{I}_G-\left(\frac{{\mathrm{\β}}_m^{\left(i\right)}}{{\mathrm{\σ}}_v^4+\mathrm{KG}{\mathrm{\σ}}_v^2{\mathrm{\β}}_m^{\left(i\right)}}\right)\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H\end{array}---\left(32\right)$

So convolution (25) and (27) can obtain for

$\begin{array}{c}{P}_{m|m,j}^{\left(i\right)}=P_{m|m-1,j}^{\left(i\right)}-P_{m|m-1,j}^{\left(i\right)}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H{\left({\left[P_{\mathrm{\υ},m,j}^{\left(i\right)}\right]}^{-1}\right)}^H\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]{\left(P_{m|m-1,j}^{\left(i\right)}\right)}^H\\ =\left(\frac{{\mathrm{\β}}_m^{\left(i\right)}{\mathrm{\σ}}_v^2}{\mathrm{KG}{\mathrm{\β}}_m^{\left(i\right)}+{\mathrm{\σ}}_v^2}\right)I_{\mathrm{KL}}={\mathrm{\δ}}_m^{\left(i\right)}{I}_{\mathrm{KL}}\end{array}---\left(33\right)$

From formula (30) and (33), getting during for diagonal matrix, with also be diagonal matrix, so formula (20) ~ (27) can be reduced to

$h_{m|m-1,j}^{\left(i\right)}={\mathrm{Fh}}_{m-1|m-1,j}^{\left(i\right)}---\left(34\right)$

${\mathrm{\β}}_m^{\left(i\right)}={\left|\mathrm{\α}\right|}^2{\mathrm{\δ}}_{m-1}^{\left(i\right)}+{\mathrm{\σ}}_w^2---\left(35\right)$

${\mathrm{\μ}}_m^{\left(i\right)}=\frac{{\mathrm{\β}}_m^{\left(i\right)}}{\mathrm{KG}{\mathrm{\β}}_m^{\left(i\right)}+{\mathrm{\σ}}_v^2}---\left(36\right)$

$h_{m|m,j}^{\left(i\right)}=(1-{\mathrm{\μ}}_m^{\left(i\right)}\mathrm{KG})h_{m|m-1,j}^{\left(i\right)}+{\mathrm{\μ}}_m^{\left(i\right)}{\left[X_m^{\left(i\right)}(p,j)\mathrm{\Ψ}(p,j)\right]}^H{y}_m^{\left(i\right)}(p,j)---\left(37\right)$

${\mathrm{\δ}}_m^{\left(i\right)}=\frac{{\mathrm{\β}}_m^{\left(i\right)}{\mathrm{\σ}}_v^2}{\mathrm{KG}{\mathrm{\β}}_m^{\left(i\right)}+{\mathrm{\σ}}_v^2}---\left(38\right)$

When employing formula (34) ~ (38) calculate, get at initial time for diagonal matrix, that just can avoid in calculating formula (25) in an iterative process is inverse, greatly reduces the amount of calculation of algorithm.

Described modification method refers to first to all carry out iterative processing, estimate the spreading code of each user, then carry out cholesky resolution process to observation data, obtain the inspection process data of each user respectively, final sampling obtains the state parameter value with maximum a posteriori probability.

Compared with prior art, the invention has the beneficial effects as follows:

(1) the CDMA multiple carrier signal with Space-Time Block Coding can be adapted under lower signal to noise ratio condition, under different spreading length, different populations and time-variant multipath channel environment, all there is higher estimated performance, and amount of calculation be lower.

(2) can estimate spreading code and information sequence simultaneously, the process of despreading and demodulation can be saved in non-cooperating communication system, greatly simplifie system.

(3) the present invention can complete the process to intermediate-freuqncy signal, also can complete the process to baseband signal or demodulation code stream simultaneously, adapts to data type extensive.

Accompanying drawing explanation

Fig. 1 is the overview flow chart of the method for the invention;

Fig. 2 is the emitting structural of MC-CDMA system;

Fig. 3 is the transmission structure of Space-Time Block Coding;

Spreading code estimated performance when Fig. 4 is different user in embodiment compares;

Information sequence estimated performance when Fig. 5 is different user in embodiment compares;

Spreading code Performance comparision when Fig. 6 is different spreading length in embodiment;

Information sequence estimated performance when Fig. 7 is different population in embodiment compares;

Fig. 8 is the tracking sampling mean of time-varying fading channels amplitude in embodiment.

Embodiment

Below in conjunction with drawings and Examples, the present invention is described in further detail.

Fig. 1 is overview flow chart of the present invention, and described in the present embodiment, method implementation procedure is as follows:

(1) receive if sampling data or base band data, estimate the spreading period of spread-spectrum signal, spreading rate and user's number, and send it to spreading code and information sequence estimation module.

(2) initialization data.To i=1 ..., I, according to B _mthe discrete uniform prior distribution initialization that (p, j) and c (p, j) obey with get ${\mathrm{\&delta;}}_0^{\left(i\right)}={\mathrm{\&epsiv;}}_i,0<{\mathrm{\&epsiv;}}_i<1,P_{0|0,j}^{\left(i\right)}={\mathrm{\&delta;}}_0^{\left(i\right)}{I}_{\mathrm{KL}},$ obeying average is 0, and covariance matrix is multiple Gaussian Profile; Make m=1, $w_0^{\left(i\right)}(p,j)=1.$

(3) for fixing first adopt the method for sequential Monte Carlo to all carry out iterative processing, obtain c _mthe estimation of (p, j), and calculate according to the thought of sequential sampling secondly adopt cholesky decomposition method to process observation data, obtain the inspection process data of each user respectively; Then according to formula (34) ~ (38), average and covariance matrix are upgraded; Finally sampling obtains the b with maximum a posteriori probability _m(p, j).

(4) importance weight is calculated

(5) double sampling.Calculate yardstick of effectively sampling wherein for the variance of importance weight, work as I _mbe less than the threshold value preset time, carry out double sampling, obtain new particle its importance weight is ${\{{\hat{w}}_m^{\left(i\right)}(p,j)=1\}}_{i=1}^I.$

(6) repeat step 3 ~ step 5 and carry out successive ignition, finally obtain each state parameter estimated result, and then spreading code and information sequence are spliced and reconstruct, if while m=200; to exit, otherwise m=m+1, enter the calculating of subsequent time.

In embodiment: number of users K is respectively 4,6,8,10, sub-carrier number N=128, each user (p=1 in p circuit-switched data, 2) spreading code is the random sequence of length G=64, the multipath way of each user is L=3, the time varying channel amplitude of fading is modeled as AR Model of First, i.e. h _m,j=Fh _{m-1, j}+ w _m, get F=0.999I _kL, w _mfor zero-mean, covariance matrix is Σ _w=0.01I _kLmultiple Gaussian Profile, adopt the STBC data of M=200 frame to process, in iterative process, get population I=100.

Fig. 4 and Fig. 5 is respectively number of users K when being 4,6,8,10, the spreading code that the present invention draws and the curve that information sequence estimated performance changes with input signal-to-noise ratio.In the multi-user case, being calculated as to be averaging the spreading code of all users and information sequence particle sample value of output signal-to-noise ratio and the error rate obtains, and the spreading code output signal-to-noise ratio in Fig. 4 is averaging rear acquisition to the result of two antennas.Embodiment result shows, when different user number, the present invention can go out spreading code and information sequence at blind estimate under comparatively Low SNR, when signal to noise ratio is greater than-5dB, when number of users is no more than 10, the spreading code average output SNR recovered is greater than 13dB, and the information sequence estimating bit error rate of all users is all lower than 10 ^-2.

Under Fig. 6 is given in different spreading length condition, the estimated performance of spreading code.As can be seen from the figure, along with the increase of G, the estimated performance of sequence is along with raising.Its main cause is in an iterative process, and the way of grouping is more, produces overlapping probability larger between signal, thus causes the performance of algorithm to reduce.

Fig. 7 gives G=64, and when population I is respectively 50,100,200, the information sequence error rate is with input signal-to-noise ratio change curve.As seen from the figure, population is larger, and its estimated performance is better, but amount of calculation is also multiplied, consideration of therefore should compromising when selecting population.

Fig. 8 gives as number of users K=4, SNR=-6dB, when iteration proceeds to m=N=200, and time-varying fading channels amplitude ss _{1, m, 1,1}tracking sampling mean.

Claims (5)

1. a CDMA multiple carrier signal spread-spectrum code and information sequence blind estimate method for simplifying, it is characterized in that: by setting up the signal model of frequency domain spread spectrum, the channel circumstance become during consideration, in conjunction with the thought of segmentation modeling, adopt the sampling particle of sequential Monte Carlo to be similar to the associating Posterior probability distribution set up, and carry out parameter state estimation according to sample value and importance weight size; Reasonably design performs flow process simultaneously, revises original iterative computation importance weight step, and makes full use of the orthogonal attributes of Space-Time Block Coding and multi-carrier modulation, the implementation procedure of method for simplifying.

2. spreading code according to claim 1 and information sequence blind estimating method, it is characterized in that, the method is mainly used in spreading code and the information sequence blind estimate of the CDMA multiple carrier signal with Space-Time Block Coding, and the channel circumstance become when signal model is considered, be modeled as the AR model of single order.

3. spreading code according to claim 1 and information sequence blind estimating method, it is characterized in that, described segmentation modeling thought refers to: Received signal strength is carried out segment processing according to different symbol ways and number of antennas, and final splicing obtains the complete spreading code of each user and information sequence.

4. spreading code according to claim 1 and information sequence blind estimating method, it is characterized in that, described simple implementation process refers to: in conjunction with the execution flow process of kalman filter method, make full use of the orthogonal attributes of Space-Time Block Coding and multi-carrier modulation, simplify iterative step, greatly reduce the computation complexity of method.

5. spreading code according to claim 1 and information sequence blind estimating method, it is characterized in that, described modification method refers to: the spreading code first estimating each user, again cholesky resolution process is carried out to observation data, obtain the inspection process data of each user respectively, final sampling obtains the state parameter value with maximum a posteriori probability.