Neural network and fluid power control

This paper will postulate the destined progress from classical to neuro-control techniques based upon the structural and functional comparison of fluid powered system with the human body. It will describe the similarities between the feed forward neural network and the proportional-integral-derivative (PID) controller. After establishing the PID controller as a limited form of feed-forward neural net, the paper will discuss the extension of using the neural network back-propagation training technique to the PID controller. Simulated results of a simple system are given to illustrate the training process. (from Authors)