GMSK调制解调的MATLAB实现

clear ,clc %main 为程序的入口 %本程序仿真GMSK信号的解调过程 %程序先调用Generator产生一个调制信号 %调制信号的抽样值传入下面的函数进行解调 %调用GMSKDecoder进行解调 %转移表，前8条路径为定值，所以全部列出 TransferTable = [ 0 +1 +1 pi/2 +1 -1 pi -1 +1 pi/2 +1 +1 0 +1 +1 pi/2 +1 -1 pi -1 +1 pi/2 +1 -1 0 +1 +1 pi/2 +1 -1 pi -1 -1 pi/2 -1 +1 0 +1 +1 pi/2 +1 -1 pi -1 -1 pi/2 -1 -1 0 +1 +1 pi/2 +1 +1 pi +1 +1 -pi/2 +1 +1 0 +1 +1 pi/2 +1 +1 pi +1 +1 -pi/2 +1 -1 0 +1 +1 pi/2 +1 +1 pi +1 -1 -pi/2 -1 +1 0 +1 +1 pi/2 +1 +1 pi +1 -1 -pi/2 -1 -1 ]; theta_now = [0 +1 +1]; %指上一个码元产生后的theta bit = [-1 -1 -1 1 1 ] %差分序列 hamming_head = zeros(8,3); %这里保存前8条路径的汉明距离 hamming_temp1 = []; hamming_temp2 = []; path = []; %路径 path_temp = []; method = 'a'; %由于每次状态的转移集合可以分成两组，所以就可以分别在码元计算之间进行交替 for i = 1:length(bit) if (i-1 > 2) %判断是否出了前8条路径，前8条路径只有3个点 disp('正在仿真8路径后的点:'); disp(i); if (i-1 > 3) %每次要调整比较的路径 if (method == 'a') method = 'b'; else method = 'a'; end end else disp('正在仿真的点数:'); disp(i); end [Sn, Cn, theta_now] = Generator(theta_now, bit(i), i-1); [path_temp,hamming_temp1,hamming_temp2] = GMSKDecoder2(Sn, Cn, i-1, method); if (i == 4) %到第四个码元时，将路径度量调整为全不为0的矩阵 %因为此前存储的hamming_head矩阵为如下格式 % 6.7432 0.5892 -7.4709 %-3.4128 7.5723 4.3539 % 0 1.7402 -9.6432 % 0 0.4883 10.9556 % 0 0 3.1054 % 0 0 -4.5701 % 0 0 -5.7368 % 0 0 2.5033 hamming_head(5:8,1) = hamming_head(2,1); hamming_head(1:4,1) = hamming_head(1,1); hamming_head(7:8,2) = hamming_head(4,2); hamming_head(5:6,2) = hamming_head(3,2); hamming_head(3:4,2) = hamming_head(2,2); hamming_head(1:2,2) = hamming_head(1,2); %现在将其调整成下面的格式 %8.9284 0.2476 -9.6936 %8.9284 0.2476 7.3007 %8.9284 6.1644 -8.2771 %8.9284 6.1644 10.5269 %0.0722 0.9976 1.4014 %0.0722 0.9976 -3.4820 %0.0722 0.5668 -8.3241 %0.0722 0.5668 5.6449 for i = 1:8 hamming(i,1) = sum(hamming_head(i,:)); %将每一行的和赋值给hamming矩阵 end %赋值后的hamming矩阵 %14.8394 %12.0642 %13.8251 %27.2332 % 4.3106 % 1.2857 %-0.2288 %-5.1446 end %每一个码元计算完成后将路径度量值赋给hamming矩阵存储 if (i-1 > 3) hamming = [hamming hamming_temp1]; end hamming_head = hamming_head + hamming_temp2; %这是前三个码元计算时用到的路径度量值存储，之后的码元GMSKDecoder2函数返回的hamming_temp2都是8*3的全零矩阵，所以相加不改变hamming_head的值 path = [path path_temp]; %路径保存 end % 开始解码 tempth = []; realpath = []; new_hamming = []; new_hamming_temp = []; %下面做的主要是根据path表把路径整理出来 for i = 1:8 tempth = path(i,1:6); temp = path(i,4:6); new_hamming_temp = hamming(i, 2); for k = 1:length(path)/6 - 1 std = path(:,1 + k*6:6 + k*6); for index = 1:8 if (temp == std(index,1:3)) tempth = [tempth std(index, 4:6)]; new_hamming_temp = [new_hamming_temp hamming(index, k + 2)]; temp = std(index, 4:6); break; end index = index + 1; end end realpath = [realpath ; tempth]; new_hamming = [new_hamming; new_hamming_temp]; end realpath2 = []; new_hamming2 = []; %整理路径度量值，使之和path相对应 for i = 1:8 temph = TransferTable(i,10:12); for k = 1:8 if temph == realpath(k,1:3) realpath2 = [realpath2;realpath(k,:)]; new_hamming2 = [new_hamming2; new_hamming(k,:)]; break; end end end finalhamming = [hamming(:,1) hamming_head new_hamming2]; for i = 1:8 finalhamming(i,1) = sum(finalhamming(i,2:length(finalhamming(1,:)))); %最终的路径度量值表 end [x y] = max(finalhamming(:,1)); path = [TransferTable(:,1:9) realpath2]; code = path(y,6:3:length(path)) %输出解码 disp('最佳路径号') disp(y)