Abstract
This paper examines an application of Neural Network for identification and control of a servovalve controlled hydraulic cylinder system. The linearized transfer function of a nonlinear plant is determined by applying a selfexcited oscillation. Neural Network is utilized on-line in mapping the inherent nonlinearities as well as in adaptive identification of system parameters. Neural Network controllers having variable proportional gain together with velocity feedback gain characteristics have been employed. The effectiveness of the Neural Network is confirmed through computer simulations and experiments.
