CN1798115A - Method of Compensating Channel Distortion in Communication System and Iterative Decision Feedback Equalizer - Google Patents

Abstract

The invention discloses a method for compensating channel distortion in a communication system and an iterative decision feedback equalizer thereof. The compensation method mainly comprises the steps of defining the length of the feedforward filter, giving an approximate solution of the feedforward filter in a discrete Fourier transform domain and the like. The iterative decision feedback equalizer mainly comprises the following components and signal flows: the received signal is output to the feedforward filter after passing through the pulse shaping filter and the sampler, the output signal of the feedforward filter is superposed with the output of the feedback filter, and then is input to the decision device after passing through the descrambler and the despreader, the hard decision of the decision device is input to the spread spectrum and the scrambler to obtain the feedback signal corresponding to each transmitting antenna and each code channel, the feedback signals of all the transmitting antennas are input to each feedback filter, the iteration is stopped after the iteration reaches the preset number, the decision device outputs the log-likelihood ratio information to the channel decoder, and the channel decoder obtains the information bit after decoding. The design method and the equalizer thereof have no error propagation phenomenon, and the realization complexity is approximate to a linear function of the channel memory length, and the method can be used for a communication system.

Description

The method and the feedback equalizer through iteration decision thereof of channel distortion in the compensation communication system

One technical field

The present invention relates to a kind of method and feedback equalizer through iteration decision of compensation communication system channel distortion that compensates the channel distortion in the communication system and eliminate the method and apparatus, particularly a kind of low complex degree of common-channel interference.

Two background technologies

In communication system, the distortion of channel will cause intersymbol interference, and receiver must take certain technology to overcome multipath fading.In second generation wireless communication system GSM, adopt viterbi equalizer to overcome multipath fading, the major defect of this technology is that implementation complexity is the exponential function of channel memory span, therefore is not suitable for wide-band communication system.In third generation wireless communication system WCDMA, time spent domain linear equalization device [was seen GLOBECOM 1999, Vol.1a, the 467-471 page or leaf, K.Hooli, M.Latva-aho, " the Multiple access interferencesuppression with linear chip equalizers in WCDMA downlink receivers " of M.Juntti] compensate for channel distortions, thereby the orthogonality between the recovery code channel, the major defect of this technology is that its implementation complexity is the cubic function of channel memory span, and its performance is subjected to the influence of residual intersymbol interference.The defective of prior art and deficiency: the technology of compensate for channel distortions commonly used has viterbi equalizer, time domain linear equalizer, DFF.The major defect of viterbi equalizer, time domain linear equalizer, DFF is the implementation complexity height, be not suitable for following broadband connections, and there is the error propagation phenomenon in decision feedback equalization.A.Burg[sees VTC autumn, in October, 2003, the 468-472 page or leaf, A.Burg, M.Rupp, S.Haene, D.Perels, N.Felber, " the Low complexityfrequency-domain equalization of MIMO channels with applications to MIMO-CDMAsystems " of W.Fichtner], [see Signals, Systems ﹠amp; Computers, The Thrity-Seventh AsilomarConference on, in November, 2003, Vol.2, the 1266-1272 page or leaf, A.Burg, M.Rupp; N.Felber, " the Practical low complexity linear equalization for MIMO-CDMAsystems " of W.Fichtner] there is capability platform in the equalizer that proposes, in real system, can't use.

Three summary of the invention

Technical problem to be solved by this invention is channel distortion and a common-channel interference in the communication system.It has overcome the major defect of prior art: complexity is at least the cubic function of channel memory span; There is the error propagation phenomenon; Or there is a capability platform.

Technical scheme of the present invention: the method for channel distortion is the design to feedback equalizer through iteration decision in the compensation communication system, and is promptly as follows to the l time iterative filter design:

(1) definition feedforward filter length L _e

(2) provide the setting of endless feedback equalizer through iteration decision.

(3) provide the approximate solution of endless feedback equalizer through iteration decision feedforward filter in discrete Fourier transform domain.

(4) feedforward filter of corresponding step 3 design, the criterion based on output Signal to Interference plus Noise Ratio maximum provides optimum feedback filter setting.

(5) according to the l time iteration output Signal to Interference plus Noise Ratio of feedback equalizer through iteration decision, calculate normalizated correlation coefficient.

(6) l ← l+1 gets back to step 1.

Realize the feedback equalizer through iteration decision of claim 1 communication system channel compensation method, it comprises: pulse shaping filter and sampler 1, feedforward filter 2, descrambler 3, despreader 4, decision device 5, spread spectrum and scrambler 6, feedback filter 7, channel decoder 8, its signal flow: after the received signal passages through which vital energy circulates is washed into mode filter and sampler 1, export to feedforward filter 2-i, the output of the output signal of feedforward filter 2-i and feedback filter 7-i is superposition mutually, remove interfering components, the signal of removing interfering components is through descrambler 3-i, behind the despreader 4-i-u, input decision device 5-i-u, the hard decision of decision device 5-i-u is behind spread spectrum and scrambler 6-i-u, obtain corresponding i transmit antennas, the feedback signal of u code channel, the feedback signal of all transmitting antennas is all imported each feedback filter 7, after iterations reaches and presets number of times, stop iteration, decision device 5 output log-likelihood ratio information are given channel decoder 8, and channel decoder 8 obtains information bit through decoding.

Beneficial effect of the present invention: do not have the error propagation phenomenon; Because utilize the fast discrete Fourier mapping algorithm to upgrade the filter setting, its implementation complexity is approximately the linear function of channel memory span.The implementation complexity that table 1 has provided equalizer compares.Wherein L is the channel memory span, and J represents observation window length, and D=L+J-1 is judgement time-delay, M _tAnd N _rBe respectively number of transmit antennas and reception antenna number.As can be seen from Table 1, the implementation complexity of time domain linear equalizer is the cubic function of channel memory span, when the channel memory span is bigger, the time domain linear equalizer can not be realized, and the implementation complexity of the feedback equalizer through iteration decision that the present invention proposes is approximately the linear function of channel memory span.And the implementation complexity of existing time domain DFF is more taller than time domain linear equalizer, and has the error propagation phenomenon in the low signal-to-noise ratio scope.

Table 1 equalizer implementation complexity

	The complex multiplication number of times
		The time domain linear equalizer	N r 2(2J+L) 2M t(2D+1)+N r 3(2J+L) 3/3
Feedback equalizer through iteration decision	M tN r(2J+L)(log2(4J+2L)-1) +N r 2M t(2J+L)(L s+1)+N r 3(L s+1)(2J+L)/3 +2M t 2N r(2J+L)(L s+1) +(2L s+1)N rM t(2J+2L)/2(log2(2J+2L)-1)

Range of application of the present invention is to overcome the multipath fading in the communication system and eliminate common-channel interference.

The implementation complexity of the feedback equalizer through iteration decision that the present invention proposes is approximately the linear function of channel memory span, and does not have error propagation phenomenon and platform effect, and the system spectrum utilance is significantly improved.

Four description of drawings

Below in conjunction with the drawings and specific embodiments the present invention is described in further detail.

Fig. 1 is the block diagram of feedback equalizer through iteration decision of the present invention.

Fig. 2 is the block diagram of feedback filter.

Fig. 3 is that 80MHz transmission bandwidth, carrier frequency 2.5GHz, maximum delay expansion 2.4us channel delay distribute power are angular distribution.

Fig. 4 is illustrated in the single-shot list and receives in the cdma system of 10MHz transmission bandwidth, and the performance of linear equalizer and feedback equalizer through iteration decision relatively.Solid line is represented the feedback equalizer through iteration decision performance curve, and dotted line is represented the linear equalizer performance curve.

Fig. 5 is illustrated in the performance that the single-shot list is received feedback equalizer through iteration decision in the cdma system of 10MHz transmission bandwidth.Solid line represents that feedback equalizer through iteration decision feedforward filter length is 66 o'clock performance curve, dotted line represents that feedback equalizer through iteration decision feedforward filter length is 82 o'clock performance curve, the performance curve when chain-dotted line is represented feedback equalizer through iteration decision feedforward filter variable-length.

Fig. 6 be illustrated in (2Tx, 2Rx) in the cdma system of 10MHz transmission bandwidth, the performance of linear equalizer and feedback equalizer through iteration decision relatively, solid line is represented the performance curve of feedback equalizer through iteration decision, dotted line is represented time domain linear equalizer performance curve.

Fig. 7 represents that the single-shot list receives the performance of feedback equalizer through iteration decision in the cdma system of 80MHz transmission bandwidth, and feedforward filter length is 576.

Fig. 8 represents (2Tx, 2Rx) performance of feedback equalizer through iteration decision in the cdma system of 80MHz transmission bandwidth.Feedforward filter length is 576.

Five embodiments

A kind of compensation communication system channel distortion methods is that it may further comprise the steps to the method for designing of the l time iterative filter of feedback equalizer through iteration decision:

(1) definition feedforward filter length L _e

(2) provide the setting of endless feedback equalizer through iteration decision;

(3) provide the approximate solution of endless feedback equalizer through iteration decision feedforward filter in discrete Fourier transform domain;

(4) feedforward filter of corresponding step 3 design, the criterion based on output Signal to Interference plus Noise Ratio maximum provides optimum feedback filter setting;

(5) according to the l time iteration output Signal to Interference plus Noise Ratio of feedback equalizer through iteration decision, calculate normalizated correlation coefficient;

(6) l ← l+1 gets back to step 1.

Compensation communication system channel distortion methods is:

The i transmit antennas feedforward filter frequency-domain expression of endless feedback equalizer through iteration decision is

$W_i^l\left[\mathrm{\ω}\right]={\left({\left(R^{l-1}\right[\mathrm{\ω}\left]\right)}^T\right)}^{-1}{\left(H_i^{*}\right[\mathrm{\ω}\left]\right)}^T---\left(1\right)$

Wherein (.) ^T, (.) ^*Represent transposition and conjugate transpose respectively.The approximate solution of endless feedforward filter in the expression formula of discrete Fourier transform domain is

$W_{\mathrm{Sub}}^l\left[f\right]={\left({\left({\hat{Q}}^l\right[f\left]\right)}^{*}{\hat{Q}}^l\right[f]+{\mathrm{\σ}}_v^2{I}_{N_r})}^{-1}\left({\left({\hat{Q}}^l\right[f\left]\right)}^{*}T\right[f\left]\right)---\left(2\right)$

Optimum feedback filter corresponding to u code channel, i transmit antennas symbol is

$b_{\mathrm{Sub}.i,l,u}^l={\mathrm{\ρ}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}{\left({\stackrel{\‾}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{\left(H\right)}_t---\left(3\right)$

$={\mathrm{\ρ}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\underset{n=1}{\overset{N_r}{\mathrm{\Σ}}}{\left({\stackrel{\‾}{W}}_{\mathrm{Sub},i,n}^l\right)}^{*}{\left(H\right)}_{(n-1)(2J+L)+1:n(2J+L),t}$

$={\mathrm{\ρ}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\underset{n=1}{\overset{N_r}{\mathrm{\Σ}}}\underset{p=1}{\overset{2J+L}{\mathrm{\Σ}}}{\left({\stackrel{\‾}{W}}_{\mathrm{Sub},i,n}^l\right)}_p^{*}{h}_{n,\mathrm{floor}((t-1)/(2D+1))+1}[L-1-t+p]$

t≠(i-1)(2D+1)+D+1

The l time, i transmit antennas iteration output Signal to Interference plus Noise Ratio are

${\mathrm{SINR}}_{\mathrm{Sub},i}^l={\left|{\left({\stackrel{\‾}{W}}_{\mathrm{Sub},j}^l\right)}^{*}{H}_{(i-1)(2D+1)+D+1}\right|}^2/(\underset{t=1,t\≠(i-1)(2D+1)+D+1}{\overset{2D+1}{\mathrm{\Σ}}}(\frac{U}{N}----\left(4\right)$

$\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\Σ}}}{\left({\mathrm{\ρ}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\right)}^2){\left|{\left({\stackrel{\‾}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_t\right|}^2+{\mathrm{\σ}}_v^2{\left|\right|{\stackrel{\‾}{W}}_{\mathrm{Sub},i}^l\left|\right|}^2)$

Next utilizes the l time, i transmit antennas iteration output Signal to Interference plus Noise Ratio calculating normalizated correlation coefficient, and the corresponding normalizated correlation coefficient of modulation system arbitrarily can calculate according to following formula

${\mathrm{\&rho;}}_{i,u}^l=E<d_{i,u}\left[n\right]{\left({\hat{d}}_{i,u}^l\right[n\left]\right)}^{*}>---\left(5\right)$

$=\underset{c,e\&Element;\mathrm{\&Pi;}}{\mathrm{\&Sigma;}}P\left\{d_{i,u}\right[n]=c,{\hat{d}}_{i,u}^l[n]=e\}{\mathrm{ce}}^{*}$

$=\underset{c,e\&Element;\mathrm{\&Pi;}}{\mathrm{\&Sigma;}}P\left\{{\hat{d}}_{i,u}^l\right[n]=e|d_{i,u}\left[n\right]=c\left\}P\right\{d_{i,u}\left[n\right]=c\}{\mathrm{ce}}^{*}$

Consider the code division multiple address communication system of MIMO, receiver as shown in Figure 1, after passages through which vital energy circulates is washed into shape and slice-level sampling 1, n _rThe received signal of root reception antenna is

$r_{n_r}\left[n\right]=\underset{m_t=1}{\overset{M_t}{\mathrm{\&Sigma;}}}\underset{l=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{h}_{m_t,n_r}\left[l\right]s_{m_t}[n-l]+v_{n_r}\left[n\right]---\left(6\right)$

Wherein transmit and be defined as

$s_{m_t}\left[n\right]=q\left[n\right]\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}\underset{k=0}{\overset{+\&infin;}{\mathrm{\&Sigma;}}}{A}_{m_t,u}{d}_{m_t,u}\left[k\right]c_u[n-\mathrm{Nk}]---\left(7\right)$

The code channel number of U for activating, d _{Mt, u}[n] m _tTransmit antennas, a u code channel c _uThe symbol of [n] carrying, it has normalized energy, and N is a spreading factor, $A_{m_t,u}=1/\sqrt{N}$ The expression power control factor, q[n] be scrambler sequence.h _{Nr, mt}[l] is m _tTransmit antennas and n _rThe channel impulse response of root reception antenna, v _Nr[n] is being added with property white Gauss noise, and variance is σ _v ², L is the channel memory span.The matrix notation of received signal is

r[n]＝Hs[n]+v[n] (8)

Wherein

$H=\left[\begin{array}{cccc}{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">H</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="H"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">H</figure-callout>\; </figure-callout>}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">H</figure-callout>}}_{1,{\mathrm{<figure-callout\; id="2"\; label="M\; t\; H"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">M</figure-callout>}}_t}& \\ H_{}{}_{\mathrm{2,1}}& {\mathrm{<figure-callout\; id="2"\; label="H"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">\; <figure-callout\; id="2"\; label="H"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">H</figure-callout>\; </figure-callout>}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; id="2"\; label="H"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">H</figure-callout>}}_{2,M_t}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ H_{N_r,1}& H_{N_r,2}& \&CenterDot;\&CenterDot;\&CenterDot;& H_{N_r,M_t}\end{array}\right]---\left(9\right)$

$s\left[n\right]=\left[s_1\right[n],\&CenterDot;\&CenterDot;\&CenterDot;,s_{M_t}[n]{]}^T---\left(10\right)$

$v\left[n\right]={\left[v_1\right[n],\&CenterDot;\&CenterDot;\&CenterDot;,v_{N_r}[n\left]\right]}^T---\left(11\right)$

$r\left[n\right]={\left[r_1\right[n],\&CenterDot;\&CenterDot;\&CenterDot;,r_{N_r}[n\left]\right]}^T---\left(12\right)$

$r_{n_r}\left[n\right]={\left[r_{n_r}\right[n-J],\&CenterDot;\&CenterDot;\&CenterDot;,r_{n_r}[n+L-1+J\left]\right]}^T---\left(13\right)$

$s_{m_t}\left[n\right]={\left[s_{m_t}\right[n-D],\&CenterDot;\&CenterDot;\&CenterDot;,s_{m_t}[n+D\left]\right]}^T---\left(14\right)$

$H_{n_r,m_t}=\left[\begin{array}{ccccc}{h}_{n_r,m_t}[L-1]& \&CenterDot;\&CenterDot;\&CenterDot;& h_{n_r,m_t}\left[0\right]& & \\ & h_{n_r,m_t}[L-1]& \&CenterDot;\&CenterDot;\&CenterDot;& h_{n_r,m_t}\left[0\right]& \\ & & & & \\ & & h_{n_r,m_t}[L-1]& \&CenterDot;\&CenterDot;\&CenterDot;& h_{n_r,m_t}\end{array}\right]---\left(15\right)$

$v_{n_r}\left[n\right]={\left[v_{n_r}\right[n-J],\&CenterDot;\&CenterDot;\&CenterDot;,v_{n_r}[n+L-1+J\left]\right]}^T---\left(16\right)$

J represents observation window length, and D=L+J-1 is the judgement time-delay.The major defect of traditional time-domain equalizer is that implementation complexity is very high and have an error propagation phenomenon, and A.Burg has proposed a kind of linear equalizer of low complex degree recently, but there is platform effect in its performance, is not suitable for being applied to practical communication system.In order to overcome platform effect and to avoid the error propagation phenomenon.According to channel condition information H, the l time iterative filter design is as follows:

(1) definition feedforward filter length L _e, general L _eBe 3 to 5 times of channel memory span.

(2) provide the setting of endless feedback equalizer through iteration decision, i transmit antennas feedforward filter frequency-domain expression is

$W_i^l\left[\mathrm{\&omega;}\right]={\left({\left(R^{l-1}\right[\mathrm{\&omega;}\left]\right)}^T\right)}^{-1}{\left(H_i^{*}\right[\mathrm{\&omega;}\left]\right)}^T---\left(17\right)$

Wherein

$R^{l-1}\left[\mathrm{\&omega;}\right]=\underset{m_t=1}{\overset{M_t}{\mathrm{\&Sigma;}}}(\frac{U}{N}-\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&rho;}}_{m_t,u}^{l-1}\right|}^2)H_{m_t}\left[\mathrm{\&omega;}\right]+{\mathrm{\&sigma;}}_v^2{I}_{N_r}---\left(18\right)$

$H_{m_t}\left[\mathrm{\&omega;}\right]={\left[{\mathrm{<figure-callout\; id="1"\; label="H"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">H</figure-callout>}}_{1,m_t}\right[\mathrm{\&omega;}],\&CenterDot;\&CenterDot;\&CenterDot;,H_{N_r,m_t}[\mathrm{\&omega;}\left]\right]}^T---\left(19\right)$

$H_{n_r,m_t}\left[\mathrm{\&omega;}\right]=\underset{l=0}{\overset{L-1}{\mathrm{\&Sigma;}}}{h}_{n_r,m_t}\left[l\right]e^{-\mathrm{j\&omega;l}}---\left(20\right)$

ρ _{Mt, u} ^L-1Be (l-1) inferior, m _tThe normalizated correlation coefficient of the remap symbol and the true symbol of transmit antennas, a u code channel.

(3) approximate solution of endless feedforward filter in the expression formula of discrete Fourier transform domain is

$W_{\mathrm{Sub}}^l\left[f\right]={\left({\left({\hat{Q}}^l\right[f\left]\right)}^{*}{\hat{Q}}^l\right[f]+{\mathrm{\&sigma;}}_v^2{I}_{N_r})}^{-1}\left({\left({\hat{Q}}^l\right[f\left]\right)}^{*}T\right[f\left]\right)---\left(21\right)$

Wherein

${\hat{Q}}^l\left[f\right]={\left[\begin{array}{cc}{\hat{H}}^l\left[2f\right]& \mathrm{\&alpha;}{\hat{H}}^l[2f+1]\end{array}\right]}^T---\left(22\right)$

$T\left[f\right]={\left[\begin{array}{cc}T\left[2f\right]I_{M_t}& T[2f+1]I_{M_t}\end{array}\right]}^T---\left(23\right)$

$T\left[f\right]=\underset{n=0}{\overset{2L_e-1}{\mathrm{\&Sigma;}}}\mathrm{\&delta;}[n-D]e^{\frac{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;nf}}{2L_e}}---\left(24\right)$

$u\left[t\right]=\left\{\begin{array}{cc}1,& 0\&le;t<L_e\\ 0,& L_e\&le;t<2L_e\end{array}\right.---\left(25\right)$

$U\left[f\right]=\underset{n=0}{\overset{2L_e-1}{\mathrm{\&Sigma;}}}u\left[t\right]e^{\frac{-\mathrm{<figure-callout\; id="2"\; label="j"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">j</figure-callout>}2\mathrm{\&pi;nf}}{2L_e}}---\left(26\right)$

α＝U[1]/U[0] (27)

${\hat{H}}^l\left[f\right]=H\left[f\right]{\mathrm{\&Sigma;}}^{l-1}---\left(28\right)$

(H[f]) _n，m＝H _n，m[f] (29)

${\mathrm{\&Sigma;}}^{l-1}=\mathrm{diag}\left(\right[\sqrt{\frac{U}{N}-\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&rho;}}_{1,u}^{l-1}\right|}^2},\&CenterDot;\&CenterDot;\&CenterDot;,\sqrt{\frac{U}{N}-\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{\left|{\mathrm{\&rho;}}_{M_t,u}^{l-1}\right|}^2}\left]\right)---\left(30\right)$

H wherein _{Nr, mt}[f] is h _{Nr, mt}The 2L of [l] _eThe discrete length Fourier transform.The feedforward filter of corresponding i transmit antennas symbol is

$W_{\mathrm{Sub},i}^l=\left[\begin{array}{c}{\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},1,i}^l\\ \&CenterDot;\\ \&CenterDot;\\ \&CenterDot;\\ {\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},N_r,i}^l\end{array}\right]---\left(31\right)$

Wherein ${\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},n_r,i}^l=\left[w_{\mathrm{Sub},n_r,i}^l\right[L_e-1],$ w _{Sub, nr, i} ^l[L _e-2], L, w _{Sub, nr, i} ^l[0]] ^*, w _{Sub, nr, i} ^l[t] is (W _Sub ^l[f]) _{Nr, i}Inverse discrete Fourier transformer inverse-discrete.

(4) feedforward filter of corresponding step 3 design based on the criterion of output Signal to Interference plus Noise Ratio maximum, corresponding to the optimum feedback filter of u code channel, i transmit antennas symbol is

$b_{\mathrm{Sub},i,t,u}^l={\mathrm{\&rho;}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{\left(H\right)}_t$

$={\mathrm{\&rho;}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\underset{n=1}{\overset{N_r}{\mathrm{\&Sigma;}}}{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},n,i}^l\right)}^{*}{\left(H\right)}_{(n-1)(2J+L)+1:n(2J+L),t}$

$={\mathrm{\&rho;}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\underset{n=1}{\overset{N_r}{\mathrm{\&Sigma;}}}\underset{p=1}{\overset{2J+L}{\mathrm{\&Sigma;}}}{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},n,i}^l\right)}_p^{*}{h}_{n,\mathrm{floor}((t-1)/(2D+1))+1}[L-1-t+p]---\left(32\right)$

t≠(i-1)(2D+1)+D+1

Wherein its integer part is got in floor (x) expression, (H) _tThe t row of representing matrix H, (H) _{(n-1) (2J+L)+and 1:n (2J+L), t}The t of representing matrix H row, from (n-1) (2J+L)+1 row formed to element of n (2J+L) row of row.

(5) utilize the l time, i transmit antennas iteration output Signal to Interference plus Noise Ratio calculating normalizated correlation coefficient.At first calculate the l time, i transmit antennas iteration output Signal to Interference plus Noise Ratio

${\mathrm{SINR}}_{\mathrm{Sub},i}^l={\left|{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},j}^l\right)}^{*}{H}_{(i-1)(2D+1)+D+1}\right|}^2/(\underset{t=1,t\&NotEqual;(i-1)(2D+1)+D+1}{\overset{2D+1}{\mathrm{\&Sigma;}}}(\frac{U}{N}-$

$\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{\left({\mathrm{\&rho;}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\right)}^2){\left|{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_t\right|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="v"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">v</figure-callout>}}^2{\left|\right|{\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\left|\right|}^2)---\left(33\right)$

Next utilizes the l time, i transmit antennas iteration output Signal to Interference plus Noise Ratio calculating normalizated correlation coefficient, and the corresponding normalizated correlation coefficient of modulation system arbitrarily can calculate according to following formula:

${\mathrm{\&rho;}}_{i,u}^l=E<d_{i,u}\left[n\right]{\left({\hat{d}}_{i,u}^l\right[n\left]\right)}^{*}>$

$=\underset{c,e\&Element;\mathrm{\&Pi;}}{\mathrm{\&Sigma;}}P\left\{d_{i,u}\right[n]=c,{\hat{d}}_{i,u}^l[n]=e\}{\mathrm{ce}}^{*}---\left(34\right)$

$=\underset{c,e\&Element;\mathrm{\&Pi;}}{\mathrm{\&Sigma;}}P\left\{{\hat{d}}_{i,u}^l\right[n]=e|d_{i,u}\left[n\right]=c\left\}P\right\{d_{i,u}\left[n\right]=c\}{\mathrm{ce}}^{*}$

Wherein П represents the modulation symbol collection, and c and e represent the point on the planisphere.The present invention provides the normalizated correlation coefficient computing formula of following QPSK and 16QAM modulation system,

Wherein

$Q\left(x\right)=\frac{1}{\sqrt{2\mathrm{\&pi;}}}\underset{x}{\overset{+\&infin;}{\&Integral;}}{e}^{-{\mathrm{<figure-callout\; id="2"\; label="t"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">t</figure-callout>}}^2/2}\mathrm{dt}---\left(36\right)$

${\mathrm{<figure-callout\; id="16"\; label="Q"\; filenames="A20041006585000212.png,A20041006585000221.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{16\mathrm{QAM},i,1}=Q\left(\sqrt{\frac{1}{5}\mathrm{SIN}{R}_{\mathrm{Sub},i}^l}\right)---\left(37\right)$

${\mathrm{<figure-callout\; id="16"\; label="Q"\; filenames="A20041006585000212.png,A20041006585000221.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{16\mathrm{QAM},i,2}=Q\left(3\sqrt{\frac{1}{5}\mathrm{SIN}{R}_{\mathrm{Sub},i}^l}\right)$

${\mathrm{<figure-callout\; id="16"\; label="Q"\; filenames="A20041006585000212.png,A20041006585000221.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{16\mathrm{QAM},i,3}=Q\left(5\sqrt{\frac{1}{5}\mathrm{SIN}{R}_{\mathrm{Sub},i}^l}\right)$

The researcher in this field can not need pay the innovation work, obtains the normalizated correlation coefficient computing formula of other modulation systems by formula (34), thereby designs the feedforward and the feedback filter of other modulation systems.From Filter Design, feedforward and feedback filter can realize that its Design of Filter complexity is O (L with fast fourier algorithm _e/ 2 (log2 (L _e)-1)) magnitude.Another very important characteristics are, decision-feedback is not made entirely true hypothesis, measure the reliability of hard decision with the normalization coherence factor, and the reliability of previous stage hard decision, be used to design feedforward filter again when prime, improve the reliability of output step by step, thereby overcome platform effect and avoided the error propagation phenomenon.Introduce the signal detection process of i transmit antennas, a u code channel below.As Fig. 1, the signal after pulse shaping and the slice-level sampling is behind feedforward filtering 2-i, and feedforward filter 2-i is output as

${\tilde{y}}_i^l\left[n\right]={\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}r\left[n\right]---\left(38\right)$

$={\mathrm{\&gamma;}}_i^l{s}_i\left[n\right]+{\tilde{v}}_i\left[n\right]$

Wherein

${\tilde{v}}_i\left[n\right]=\underset{t=1,t\&NotEqual;{(t-1)}^{*}(2D+1)+D+1}{\overset{M_t(2D+1)}{\mathrm{\&Sigma;}}}{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_t{s}_{\mathrm{floor}((t-1)/(2D+1))+1}[n-D+\mathrm{mod}((t-1),2D+1)]---\left(39\right)$

$+{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}v\left[n\right]$

${\mathrm{\&gamma;}}_i^l={\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_{(i-1)(2D+1)+D+1}---\left(40\right)$

s _i[n] is the signal of i root antenna emission.Corresponding to m _tThe decision-feedback signal of transmit antennas, a u code channel can be by decision device 5-m _t-u is reconstructed into

${\hat{s}}_{m_t,u}^{l-1}\left[n\right]=q\left[n\right]\underset{k=0}{\overset{+\&infin;}{\mathrm{\&Sigma;}}}A{\hat{d}}_{m_t,u}^{l-1}\left[k\right]c_u[n-\mathrm{Nk}]---\left(41\right)$

Wherein

Be m _tTransmit antennas, a u code channel, (l-1) level, the hard decision in a k moment.As Fig. 2, corresponding m _tThe feedback signal input feedback filter 7-i-m of transmit antennas _t1 or feedback filter 7-i-m _t2, and according to modulation system selection feedback filter.The output of feedback filter 7-i and feedforward filter 2-i output superposition mutually can get,

${\stackrel{\&OverBar;}{y}}_i^l\left[n\right]={\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}r\left[n\right]---\left(42\right)$

$-\underset{t=1,t\&NotEqual;{(i-1)}^{*}(2D+1)+D+1}{\overset{M_t(2D+)}{\mathrm{\&Sigma;}}}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{b}_{i,t,u}^l{\hat{s}}_{\mathrm{floor}((k-1)/(2D+1))+1,u}^{l-1}[n-D+\mathrm{mod}((t-1),2D+1)]$

$={\mathrm{\&gamma;}}_i^l{s}_i\left[n\right]+{\stackrel{\&OverBar;}{v}}_i\left[n\right]$

Wherein

${\stackrel{\&OverBar;}{v}}_i^l\left[n\right]=\underset{t=1,t\&NotEqual;{(i-1)}^{*}(2D+1)+D+1}{\overset{M_t(2D+1)}{\mathrm{\&Sigma;}}}{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_r{s}_{\mathrm{floor}((t-1)/(2D+1))+1}[n-D+\mathrm{mod}((t-1),2D+1)]---\left(43\right)$

$-\underset{t=1,t\&NotEqual;{(i-1)}^{*}(2D+1)+D+1}{\overset{M_t(2D+)}{\mathrm{\&Sigma;}}}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{b}_{i,t,u}^l{\hat{s}}_{\mathrm{floor}((k-1)/(2D+1))+1,u}^{l-1}[n-D+\mathrm{mod}((t-1),2D+1)]$

$+{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}v\left[n\right]$

Residual interference plus noise v _i ^lThe power of [n] is

$P_{{\stackrel{\&OverBar;}{v}}_i^l}=\underset{t=1,t\&NotEqual;(i-1)(2D+1)+D+1}{\overset{2D+1}{\mathrm{\&Sigma;}}}(\frac{U}{N}-\frac{1}{N}\underset{u=1}{\overset{U}{\mathrm{\&Sigma;}}}{\left({\mathrm{\&rho;}}_{\mathrm{floor}((t-1)/(2D+1))+1,u}^{l-1}\right)}^2)|{\left({\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\right)}^{*}{H}_t{|}^2+{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="v"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">v</figure-callout>}}^2{\left|\right|{\stackrel{\&OverBar;}{W}}_{\mathrm{Sub},i}^l\left|\right|}^2)---\left(44\right)$

In above-mentioned derivation, use following hypothesis

$E<{\mathrm{<figure-callout\; id="1"\; label="d\; i"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">d</figure-callout>}}_{i_{}}\left[n_1\right]{\left({\hat{d}}_{{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">i</figure-callout>}}_2,u_2}^{l-1}\right[n_2\left]\right)}^{*}>={\mathrm{\&rho;}}_{i_1,u_1}^{l-1}\mathrm{\&delta;}[n_1-n_2]\mathrm{\&delta;}[u_1-u_2]\mathrm{\&delta;}[i_1-i_2]---\left(45\right)$

$E<{\hat{d}}_{{\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">i</figure-callout>}}_1,u_1}\left[n_1\right]{\left({\hat{d}}_{{\mathrm{<figure-callout\; id="2"\; label="i"\; filenames="A20041006585000201.png"\; state="\{\{state\}\}">i</figure-callout>}}_2,u_2}^{l-1}\right[n_2\left]\right)}^{*}>=\mathrm{\&delta;}[n_1-n_2]\mathrm{\&delta;}[u_1-u_2]\mathrm{\&delta;}[i_1-i_2]---\left(46\right)$

$E<{\hat{d}}_{{\mathrm{<figure-callout\; id="1"\; label="i"\; filenames="A20041006585000201.png,A20041006585000212.png"\; state="\{\{state\}\}">i</figure-callout>}}_1,u_1}^{l-1}\left[n_1\right]{\left(v_{n_r}\right[n_2\left]\right)}^{*}>=0---\left(47\right)$

Through after descrambling 3 and the despreading 4, be corresponding to the symbol judgement variable of i transmit antennas, a u code channel

${\stackrel{\&OverBar;}{d}}^l\left[k\right]=\frac{1}{\sqrt{N}}\underset{n=\mathrm{Nk}}{\overset{N(k+1)-1}{\mathrm{\&Sigma;}}}{c}_u[n-\mathrm{Nk}]{\left(q\right[n\left]\right)}^{*}{\stackrel{\&OverBar;}{y}}_i^l\left[n\right]---\left(48\right)$

$={\mathrm{\&gamma;}}_i^l{d}_{i,u}\left[k\right]+\frac{1}{\sqrt{N}}\underset{n=\mathrm{Nk}}{\overset{N(k+1)-1}{\mathrm{\&Sigma;}}}{c}_u[n-\mathrm{Nk}]{\left(q\right[n\left]\right)}^{*}{\stackrel{\&OverBar;}{v}}_i^l\left[n\right]$

Can obtain the hard decision and the log-likelihood ratio information of i transmit antennas, a u code channel according to formula (48).Should be noted that especially above-mentioned filtering can realize with fast fourier transformation algorithm, so the implementation complexity of filtering is O (L _e/ 2 (log2 (L _e)-1)).Initial level can adopt the output of the equalizer of A.Burg proposition.The present invention only provides based on parallel feedback equalizer through iteration decision method and setting, and the researcher in this field can not need pay the innovation work, provides feedback equalizer through iteration decision method and setting based on serial and grouping.Table 2 has provided the cdma system parameter, and table 3 and Fig. 3 have provided 10MHz and 80MHz transmission bandwidth channel delay distribute power respectively.

Table 2 system parameters

Carrier frequency	2.5GHz
		Maximum Doppler frequency-shift	69.4Hz
Spreading code	Walsh code, spreading factor 8
		Transmission bandwidth	10MHz & 80MHz
Channel estimating	Desirable
		Modulation system	16QAM

Table 3 10MHz transmission bandwidth, carrier frequency 2.5GHz channel delay distribute power

The tap sequence number	Postpone [ns]	Power [dB]
			1	0	-3.2
2	100	-5
			3	400	-4.5
4	500	-3.6
			5	700	-3.9
6	800	0
			7	950	-3
8	1200	-1.2
			9	1350	-5
10	1450	-3.5

Realize the feedback equalizer through iteration decision of claim 1 compensation communication system channel distortion methods, it comprises: pulse shaping filter and sampler 1, feedforward filter 2, descrambler 3, despreader 4, decision device 5, spread spectrum and scrambler 6, feedback filter 7, channel decoder 8, its signal flow: after the received signal passages through which vital energy circulates is washed into mode filter and sampler 1, export to feedforward filter 2-i, the output of the output signal of feedforward filter 2-i and feedback filter 7-i is superposition mutually, remove interfering components, the signal of removing interfering components is through descrambler 3-i, behind the despreader 4-i-u, input decision device 5-i-u, the hard decision of decision device 5-i-u is behind spread spectrum and scrambler 6-i-u, obtain corresponding i transmit antennas, the feedback signal of u code channel, the feedback signal of all transmitting antennas is all imported each feedback filter 7, after iterations reaches and presets number of times, stop iteration, decision device 5 output log-likelihood ratio information are given channel decoder 8, and channel decoder 8 obtains information bit through decoding.

The signal flow of feedback filter 7-i is as follows: m _tThe feedback signal input feedback filter 7-i-m of transmit antennas _t1 or feedback filter 7-i-m _t2, and according to modulation system selection feedback filter 7-i-m _t1 or feedback filter 7-i-m _t2, all feedback filter { 7-i-m _t1,7-i-m _t2}m _t=1 M _tOutput phase superposition, obtaining feedback filter 7-i must export, this output and the output of feedforward filter 2-i is superposition mutually.

All devices of feedback equalizer through iteration decision are known, and those skilled in the art can realize the present invention according to design procedure and signal flow.

In conjunction with Fig. 4,5,6,7,8, following performance relatively in, cdma communication system always is in full load condition, promptly all code channels all are activated.The performance that Fig. 4 provides time domain linear equalizer and feedback equalizer through iteration decision compares.The feedforward filter length of the filter length of time domain linear equalizer and feedback equalizer through iteration decision all is 50, and can calculate feedback filter length from equation (18) is 64.As can be seen from Figure 4 at initial level (equalizer that A.Burg proposes) cisco unity malfunction, the feedback equalizer through iteration decision that the present invention proposes is handled through second iteration can obtain satisfactory performance, if target bit is 0.03, compare with the time domain linear equalizer, feedback equalizer through iteration decision can be obtained the 4.5dB performance gain.Because the initial level poor-performing, there is capability platform in feedback equalizer through iteration decision in bit error rate 0.01, and in not adding the communication system of coding, most interested bit error rate scope is between 0.1 to 0.01 fortunately.Receive in the cdma system of 10MHz transmission bandwidth at the single-shot list, the performance of linear equalizer and feedback equalizer through iteration decision relatively.Solid line is represented the feedback equalizer through iteration decision performance curve, and dotted line is represented the linear equalizer performance curve.

Fig. 5 provides the single-shot list and receives in the cdma system of 10MHz transmission bandwidth, the performance of different length feedback equalizer through iteration decision.Solid line represents that feedback equalizer through iteration decision feedforward filter length is 66 o'clock performance curve, dotted line represents that feedback equalizer through iteration decision feedforward filter length is 82 o'clock performance curve, the performance curve when chain-dotted line is represented feedback equalizer through iteration decision feedforward filter variable-length.Increasing the initial level filter length as can be seen from Figure 5 can eliminate capability platform and improve performance.In order to reduce implementation complexity, can adopt the feedback equalizer through iteration decision of feedforward filter length 66, to compare with the feedback equalizer through iteration decision of feedforward filter length 82, its performance loss that causes can be ignored.

Fig. 6 provide (2Tx, 2Rx) in the cdma system of 10MHz transmission bandwidth, the performance of linear equalizer and feedback equalizer through iteration decision relatively, solid line is represented the performance curve of feedback equalizer through iteration decision, dotted line is represented time domain linear equalizer performance curve.Being set to of feedback equalizer through iteration decision: in initial level feedforward filter length is 164, and later feedforward filter length at different levels are 100.The time domain linear equalizer filter lengths is 100.More as can be seen, feedback equalizer through iteration decision can be obtained the satisfactory performance gain from performance.

Fig. 7 and Fig. 8 have provided the performance of feedback equalizer through iteration decision when transmission bandwidth is 80MHz.The channel memory span is 192, because the implementation complexity of time domain linear equalizer is the cube of channel memory span, is difficult to realize that therefore the end provides the performance of time domain linear equalizer on engineering.Feedback equalizer through iteration decision can be obtained satisfactory performance as can be seen from Figures 7 and 8.Fig. 7 is given in the performance that the single-shot list is received feedback equalizer through iteration decision in the cdma system of 80MHz transmission bandwidth.Feedforward filter length is 576.Fig. 8 (2Tx, 2Rx) among the CDMA of 80MHz transmission bandwidth the performance of feedback equalizer through iteration decision as can be seen from Figures 7 and 8 feedback equalizer through iteration decision can obtain satisfactory performance.

Key problem in technology point of the present invention and desire protection point: feedback equalizer through iteration decision is the approximate solution of endless section feedback equalizer through iteration decision.In each level, after limit for length's feedforward filter is given, feedback filter makes residual interference power minimum, calculates the normalization coherence factor by the output signal-to-noise ratio of equalizer, and current normalization coherence factor is used for upgrading limit for length's feedforward filter and feedback filter.The realization of filter can utilize the fast discrete Fourier mapping algorithm to realize.

Claims (4)

1, a kind of communication system channel distortion compensating method, is to the design method of the 1st iterative filter of iterative decision feedback equalizer, it comprises the following steps: (1) Define the feedforward filter length L _e ; (2) Provide the settings of the infinitely long iterative decision feedback equalizer; (3) Give the approximate solution of the feedforward filter of the infinite-length iterative decision feedback equalizer in the discrete Fourier transform domain; (4) Corresponding to the feed-forward filter designed in step 3, based on the maximum criterion of the output signal-to-interference-noise ratio, the optimal feedback filter setting is given; (5) Calculate the normalized correlation coefficient according to the 1st iteration output SINR of the iterative decision feedback equalizer; (6) l←l+1, return to step 1. 2. The communication system channel distortion compensation method according to claim 1, characterized in that: The frequency domain expression of the feedforward filter of the i-th transmit antenna of the infinite-length iterative decision feedback equalizer is ${\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; R</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; T</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; \&omega;</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; T</span>}}$ The approximate solution of the infinite feed-forward filter is expressed in the discrete Fourier transform domain as ${\mathrm{<span\; class="notranslate">\; W</span>}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; (</span>{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{\mathrm{<span\; class="notranslate">\; I</span>}}_{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; (</span>{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; Q</span>}}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}\mathrm{<span\; class="notranslate">\; T</span>}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; )</span>$ The optimal feedback filter corresponding to the u-th code channel and the i-th transmit antenna symbol is ${\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; )</span>}_{\mathrm{<span\; class="notranslate">\; t</span>}}$ $<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; h</span>}<span\; class="notranslate">\; )</span>}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; :</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}}$ $<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{{\mathrm{<span\; class="notranslate">\; N</span>}}_{\mathrm{<span\; class="notranslate">\; r</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; J</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; L</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; no</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}_{\mathrm{<span\; class="notranslate">\; p</span>}}^{<span\; class="notranslate">\; *</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; L</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; ]</span>$ t≠(i-1)(2D+1)+D+1 The output SINR of the lth and i-th transmitting antenna iterations is ${\mathrm{<span\; class="notranslate">\; SINR</span>}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; =</span>{<span\; class="notranslate">\; |</span>{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}_{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\underset{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \&NotEqual;</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>\frac{\mathrm{<span\; class="notranslate">\; u</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}<span\; class="notranslate">\; -</span>$ $\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; N</span>}}\underset{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; u</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}<span\; class="notranslate">\; (</span>{{\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{\mathrm{<span\; class="notranslate">\; floor</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; /</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; D.</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>{<span\; class="notranslate">\; |</span>{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}{\mathrm{<span\; class="notranslate">\; h</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; \&sigma;</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}^{\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; W</span>}}}_{\mathrm{<span\; class="notranslate">\; Sub</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; |</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; )</span>$ Secondly, the normalized correlation coefficient is calculated by using the iterative output signal-to-interference-noise ratio of the lth and i-th transmitting antennas, and the normalized correlation coefficient corresponding to any modulation method can be calculated according to the following formula ${\mathrm{<span\; class="notranslate">\; \&rho;</span>}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; <</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span>{<span\; class="notranslate">\; (</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; )</span>}^{<span\; class="notranslate">\; *</span>}<span\; class="notranslate">\; ></span>$ $<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; ,</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; ce</span>}}^{<span\; class="notranslate">\; *</span>}$ $<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; \&Pi;</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; \{</span>{\stackrel{<span\; class="notranslate">\; ^</span>}{\mathrm{<span\; class="notranslate">\; d</span>}}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}^{\mathrm{<span\; class="notranslate">\; l</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; \}</span>\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; \{</span>{\mathrm{<span\; class="notranslate">\; d</span>}}_{{\mathrm{<span\; class="notranslate">\; m</span>}}_{\mathrm{<span\; class="notranslate">\; t</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; [</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; ]</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; \}</span>{\mathrm{<span\; class="notranslate">\; ce</span>}}^{<span\; class="notranslate">\; *</span>}$ 3. Realize the iterative decision feedback equalizer of claim 1 communication system channel distortion compensation method, it comprises: Pulse shaping filter and sampler [1], also comprise feed-forward filter [2], descrambler [3], Despreader [4], decider [5], spread spectrum and scrambler [6], feedback filter [7], channel decoder [8], characterized in signal flow: the received signal is filtered by pulse shaping After the filter and sampler [1], the output is sent to the feedforward filter [2-i], and the output signal of the feedforward filter [2-i] is superimposed with the output of the feedback filter [7-i] to remove interference Components, the signal from which the interference component has been removed passes through the descrambler [3-i] and the despreader [4-i-u], and then enters the decision unit [5-i-u], and the hard decision of the decision unit [5-i-u] is spread spectrum and After the scrambler [6-i-u], the feedback signal corresponding to the i-th transmitting antenna and the u-th code channel is obtained, and the feedback signals of all transmitting antennas are input into each feedback filter [7], and the number of iterations reaches the preset number After that, the iteration is stopped, and the decision unit [5] outputs the log likelihood ratio information to the channel decoder [8], and the channel decoder [8] decodes to obtain information bits. 4. The iterative decision feedback equalizer for realizing the channel distortion compensation method in the communication system of claim 1 according to claim 2, characterized in that the signal flow of the feedback filter [7-i] is as _follows : The feedback signal is input to the feedback filter [7-im _t 1] or the feedback filter [7-im _t 2], and the feedback filter [7-im _t 1] or the feedback filter [7-im _t 2] is selected according to the modulation mode ], all feedback filters ${\{7-i-m_t\mathrm{1,7}-i-m_{t}2\}}_{m_t=1}^{m_t}$ The output of the feedback filter [7-i] is superimposed to obtain the output of the feedback filter [7-i], which is superimposed with the output of the feedforward filter [2-i].

Images