matlab-(含教程)基于V-BLAST的ML检测算法MATLAB仿真

clc; clear; close all; warning off; addpath(genpath(pwd)); Tx_n = 4; %发送天线的数目 Rx_n = 4; %接收天线的数目 N_DLRB = 5; %资源块个数 N_OFDM = 7; %OFDM符号个数 a_length = 6720; %比特流的长度为3360 snr_dB=-5:2:15; %信噪比的取值情况 for i=1:length (snr_dB) snr(i)=10^(snr_dB(i)/10); blast_error_bit=0; mlblast_error_bit=0; mlblast1_error_bit=0; mlblast2_error_bit=0; for numm=1:2 %每一个SNR值计算多次，以增大数据量 a=rand(1,a_length); for L=1:a_length if a(L)>0.5 a(L)=1; else a(L)=0; end end %产生比特流 a_mod=QAM(a_length,a); %进行调制,采用16QAM的方式 a_=reshape(a_mod,Tx_n,length(a_mod)/Tx_n); %映射至四个天线端口 m=1; %信号检测后输出信号的序列号，在这里对其进行初始化 for ll=1:N_OFDM for kk=N_DLRB*12:-1:1 a1(kk,ll)=a_(1,(ll-1)*N_DLRB*12+abs(kk-N_DLRB*12)+1); a2(kk,ll)=a_(2,(ll-1)*N_DLRB*12+abs(kk-N_DLRB*12)+1); a3(kk,ll)=a_(3,(ll-1)*N_DLRB*12+abs(kk-N_DLRB*12)+1); a4(kk,ll)=a_(4,(ll-1)*N_DLRB*12+abs(kk-N_DLRB*12)+1); %进行资源映射 a_all=[a1(kk,ll),a2(kk,ll),a3(kk,ll),a4(kk,ll)].'; %取出相同时频位置上的资源粒子，编为一组，进行信号检测 sigma = 1/sqrt(2*Rx_n*snr(i)); %噪声的标准差，假设信号的能量为1 awgn_noise = sigma/sqrt(2)*(randn(Rx_n,1)+j*randn(Rx_n,1)); %产生高斯白噪声 H=(randn(Rx_n,Tx_n)+j*randn(Rx_n,Tx_n))/sqrt(2); %信道冲击响应矩阵 r=H*a_all+awgn_noise ; %相同时频位置上接收到的数据 %基于ZF的排序V-BLAST算法 H=(randn(Rx_n,Tx_n)+j*randn(Rx_n,Tx_n))/sqrt(2); r=H*a_all+awgn_noise; %相同时频位置上接收到的数据 mm=1:Tx_n; for ii=1:Tx_n G3=pinv(H); %滤波矩阵 G_=sum(abs(G3).^2,2); [m_,ki]=min(G_); %找出功率最小的一行,该行的可靠性最高 Z=G3(ki,:)*r; dec5(mm(ki),m)=quan_16QAM_1(Z); %量化判决过程 r=r-H(:,ki)*dec5(mm(ki),m); H(:,ki)=[]; mm(ki)=[]; end %基于ZF的有所保留的排序V-BLAST算法(1层，4个取值） mlblast_16=mlblast_qam_1(m,a_all,awgn_noise,Rx_n,Tx_n); dec6(:,m)=mlblast_16; %基于ZF的有所保留的排序V-BLAST算法(1层，2个取值） mlblast1_16=mlblast_qam_2(m,a_all,awgn_noise,Rx_n,Tx_n); dec7(:,m)=mlblast1_16; %基于ZF的有所保留的排序V-BLAST算法(1层，6个取值） mlblast2_16=mlblast_qam_3(m,a_all,awgn_noise,Rx_n,Tx_n); dec8(:,m)=mlblast2_16; m=m+1; end end dec5_=reshape(dec5,1,length(a_mod)); dec_blast_zf2=demod_QAM(dec5_); %解调过程，对应于QPSK dec6__=reshape(dec6,1,length(a_mod)); dec_mlblast=demod_QAM(dec6__); %解调过程，对应于QPSK dec7__=reshape(dec7,1,length(a_mod)); dec_mlblast1=demod_QAM(dec7__); %解调过程，对应于QPSK dec8__=reshape(dec8,1,length(a_mod)); dec_mlblast2=demod_QAM(dec8__); %解调过程，对应于QPSK err_num_blast_zf2=error_count(a,dec_blast_zf2); %统计基于MMSE的V-BLAST算法下错误比特个数 err_num_mlblast=error_count(a,dec_mlblast); %统计ZF算法下错误比特个数 err_num_mlblast1=error_count(a,dec_mlblast1); %统计ZF算法下错误比特个数 err_num_mlblast2=error_count(a,dec_mlblast2); %统计ZF算法下错误比特个数 blast_error_bit=blast_error_bit+err_num_blast_zf2; mlblast_error_bit=mlblast_error_bit+err_num_mlblast; mlblast1_error_bit=mlblast1_error_bit+err_num_mlblast1; mlblast2_error_bit=mlblast2_error_bit+err_num_mlblast2; end err_ber_blast_zf2(i)=blast_error_bit/(length(a)*numm) %统计BLAST算法下错误比特率 err_ber_mlblast(i)=mlblast_error_bit/(length(a)*numm) err_ber_mlblast1(i)=mlblast1_error_bit/(length(a)*numm) err_ber_mlblast2(i)=mlblast2_error_bit/(length(a)*numm) end semilogy (snr_dB,err_ber_blast_zf2,'r-x') hold on semilogy (snr_dB,err_ber_mlblast,'b-*') semilogy (snr_dB,err_ber_mlblast1,'r:+') semilogy (snr_dB,err_ber_mlblast2,'g-o') grid on; axis([-5,15,0.0001,1]); legend('vblast','ml-blast'); xlabel('Eb/No in dB'); ylabel('BER');