BP Neural Network-based Parameter Self-Learning Control

Parameter self-learning control based on BP neural network involves: (1) defining the BP network structure by setting input layer nodes M and hidden layer nodes Q, initializing weight coefficients for each layer, and selecting learning rate and momentum factor with k=1; (2) sampling reference input rin(k) and output yout(k) to calculate error error(k)=rin(k)-yout(k); (3) computing inputs/outputs of NN layers where output layer yields three adjustable PID parameters; (4) calculating PID controller output u(k) using equation (3.34); (5) performing NN training to adaptively adjust PID parameters via online weight updates.