BP神经网络PID控制器

function [sys,x0,str,ts,simStateCompliance] = exppidf(t,x,u,flag) switch flag, case 0, [sys,x0,str,ts,simStateCompliance]=mdlInitializeSizes; case 1, sys=mdlDerivatives(t,x,u); case 2, sys=mdlUpdate(t,x,u); case 3, sys=mdlOutputs(t,x,u); case 4, sys=mdlGetTimeOfNextVarHit(t,x,u); case 9, sys=mdlTerminate(t,x,u); otherwise DAStudio.error('Simulink:blocks:unhandledFlag', num2str(flag)); end function [sys,x0,str,ts,simStateCompliance]=mdlInitializeSizes sizes=simsizes; sizes.NumContStates=0; sizes.NumDiscStates=3; sizes.NumOutputs=4; sizes.NumInputs=7; sizes.DirFeedthrough=1; sizes.NumSampleTimes=1; sys=simsizes(sizes); x0=[1,0.1,10]; str=[]; ts=[0 0]; simStateCompliance = 'UnknownSimState'; function sys=mdlDerivatives(t,x,u) sys = [x]; function sys=mdlUpdate(t,x,u) T=0.1; x=[u(5);x(2)+u(5)*T;(u(5)-u(4))/T];%e(k),e(k)求和,e(k)求微分 sys=[x(1);x(2);x(3)]; function sys=mdlOutputs(t,x,u) xite=0.2; alfa=0.05; IN=3;H=5;OUT=3; T=0.1; global wi; global wo; global wi_1; global wi_2; global wi_3; global wo_1; global wo_2; global wo_3; global Kp; global Ki; global Kd; %wi=rand(5,3);wo=rand(3,5); Oh=zeros(1,5); I=Oh; K1=zeros(3,1);K=zeros(1,3);dyu=0;dK=zeros(1,3);delta3=zeros(1,3); d_wo=zeros(3,5);segma=zeros(1,5);delta2=zeros(1,5);d_wi=zeros(5,3);dO=zeros(1,5); xi=[u(1),u(3),u(5)]; epid=[x(1);x(2);x(3)];%用于传给pid控制器的误差 I=xi*wi';%隐含层的输入 for j=1:1:5 Oh(j)=(exp(I(j))-exp(-I(j)))/(exp(I(j))+exp(-I(j)));%隐含层的输出 end K1=wo*Oh'; %输出层的输入 for i=1:1:3 K(i)=exp(K1(i))/(exp(K1(i))+exp(-K(i)));%输出层的输出 end u_k=K*epid; %以下是权值调整 %隐含层至输出层的权值调整 dyu=sign((u(3)-u(2))/(u(7)-u(6)+0.01)); for j=1:1:3 dK(j)=2/(exp(K1(j))+exp(-K1(j)))^2; %输出层的输出的一阶导 end for i=1:1:3 delta3(i)=u(5)*dyu*epid(i)*dK(i); %输出层的delta end for j=1:1:3 for i=1:1:5 d_wo=xite*delta3(j)*Oh(i)+alfa*(wo_1-wo_2); end end wo=wo_1+d_wo+alfa*(wo_1-wo_2); %以下是输入层至隐含层的权值调整 for i=1:1:5 dO(i)=4/(exp(I(i))+exp(-I(i)))^2; end segma=delta3*wo; for i=1:1:5 delta2(i)=dO(i)*segma(i); end %for i=1:1:5 % for j=1:1:3 % d_wi=xite*delta2(i)*xi(j)+alfa*(wi_1-wi_2); % end %end d_wi=xite*delta2'*xi; wi=wi_1+d_wi+alfa*(wi_1-wi_2); wo_3=wo_2; wo_2=wo_1; wo_1=wo; wi_3=wi_2; wi_2=wi_1; wi_1=wi; Kp=K(1);Ki=K(2);Kd=K(3); sys=[u_k,Kp,Ki,Kd]; %sys=[1,2,3,4]; function sys=mdlGetTimeOfNextVarHit(t,x,u) sampleTime = 0.1; % Example, set the next hit to be one second later. sys = t + sampleTime; function sys=mdlTerminate(t,x,u) sys = [];