Self-Tuning PID Controller based on Recurrent Neural Networks

R. Jafari and R. Dhaouadi (UAE)


Recurrent neural network, Adaptive control, PID controller, Nonlinear system.


In this paper a new self-tuning PID control scheme for a nonlinear system is proposed. The controller tuning is based on Recurrent Neural Networks. The adaptive control scheme uses a model reference technique to make the position of a single degree of freedom pendulum tracks the desired position. The specialized Real Time Recurrent Learning algorithm was used to train the network. The advantages of the proposed technique are demonstrated through simulation, investigating the dynamic modeling properties of recurrent neural networks and their adaptive ability to control a nonlinear pendulum system.

Important Links:

Go Back