Using computational intelligence strategies in delay compensation for real-time systems

This paper presents an approach for realtime systems, as hybrid testing, active and semiactive control, in which time delay compensation is implemented using adaptive prediction by means of neural networks which provides a new command signal to compensate the actuator (or device) delay. In presence of noisy signals, linear neural networks demonstrate much more capacity and robustness than other common methods for prediction. They provide a smoother signal avoiding the slight discontinuities which can be found in other methods in real cases when noise is present.