锐客网

  1. 首页 > 睿知 > it技术 > >

智能控制算法|MATLAB进行神经网络之单神经元自适应控制

智能控制算法算法迭代加深

MATLAB实现神经网络之单神经元自适应控制 【智能控制算法|MATLAB进行神经网络之单神经元自适应控制】系统对象
y(k) = 0.368y(k-1)+0.26y(k-2)+0.1u(k-1)+0.632u(k-2)
系统输入
rin(k)=0.5sgn(sin(4πt))

clc clear%定义输入矩阵 x = [0,0,0]'; %定义学习速率 xite1 = 0.4; xite2 = 0.3; xite3 = 0.5; %初始化权重 wk1_1 = 0.1; wk2_1 = 0.1; wk3_1 = 0.1; %初始化误差 erro_1 = 0; erro_2 = 0; %初始化被控对象模型 y_1 = 0; y_2 = 0; y_3 = 0; u_1 = 0; u_2 = 0; u_3 = 0; %定义时间步长 ts = 0.001; %迭代学习 for k=1:1:1000 %循环迭代次数 time(k) = k*ts; %时间为K*ts总时间也就是1 yd(k) = 0.5*sign(sin(4*pi*time(k))); %目标期望值 y(k) = 0.368*y_1+0.26*y_2+0.1*u_1+0.632*u_2; %被控对象动态模型模型 erro(k) = yd(k)-y(k); %获取误差%更新权重 wk1(k) = wk1_1+xite1*erro(k)*u_1*x(1); wk2(k) = wk2_1+xite2*erro(k)*u_1*x(2); wk3(k) = wk3_1+xite3*erro(k)*u_1*x(3); %定义比例系数 K = 0.12; %给输入矩阵赋值 x(1) = erro(k)-erro_1; x(2) = erro(k); x(3) = erro(k)-2*erro_1+erro_2; %将权重求和 wadd(k) = abs(wk1(k))+abs(wk2(k))+abs(wk3(k)); w11(k) = wk1(k)/wadd(k); %得w11 w22(k) = wk2(k)/wadd(k); %w22 w33(k) = wk3(k)/wadd(k); %w33 w = [w11(k),w22(k),w33(k)]; %增量 u(k) = u_1 + K*w*x; erro_2 = erro_1; %e(k-2) erro_1 = erro(k); %e(k-1)u_3 = u_2; u_2 = u_1; u_1 = u(k); y_3 = y_2; y_2 = y_1; y_1 = y(k); wk1_1 = wk1(k); %w1(k-1) wk3_1 = wk3(k); wk2_1 = wk2(k); end%---------------结果可视化-------------------%%显示被控模型曲线和设定曲线 figure(1); plot(time,yd,'r',time,y,'k','linewidth',2); xlabel('time(s)'); ylabel('yd,y'); legend('ideal position','position tracking'); title('被控模型曲线和设定曲线')%显示误差 figure(2); plot(time,erro,'r','linewidth',2); xlabel('time(s)'); ylabel('erro(t)'); title('误差曲线') %显示控制u（t）输出曲线 figure(3); plot(time,u,'r','linewidth',2); xlabel('time(s)'); ylabel('u(t)'); title('控制量u（t）输出曲线')%显示w参数调整曲线 figure(4); plot(time,w11,'g',time,w22,'r',time,w33,'b','linewidth',2); xlabel('time(s)'); ylabel('w(t)'); legend('w11(t)','w22(t)','w33(t)'); title('w11,w22,w33参数调整曲线')

    推荐阅读


    上一篇：matlab|如何用matlab画一个球

    下一篇：vue|一篇文章汇总git常用命令