xhmu.m

在matlab下实现系统辨识设计

x  = randn(1,700);                  %作为系统的输入
       b  = fir1(63,0.5)                    %确定未知系统
       n  = 0.1*randn(1,700);               %加入噪声
       d  = filter(b,1,x)+n;                  %期望信号
       f  =zeros(1,320),t=1;
     for mu=0:0.0001:0.032                  % 循环步长mu值采样
         h = adaptfilt.lms(64,mu);
      [y,e] = filter(h,x,d);
      f(1,t)=mean(abs(e));                   %对采样点取绝对值的均值
           t=t+1;
       end
       t=1;  
       subplot(1,1,1);stem([f]);axis([0,320,-0.1,1]); 
       title('e-mu关系图');xlabel('步长采样点'); ylabel('误差值');