Neuro Controlled Active Suspension with Preview for Ride Comfort

The objective of this study is to develop a neuro controlled active suspension for the ride quality improvement. The performance index of the optimal control is represented as the frequency-shaped using Parseval's theorem. The incorporation of frequency-dependent weighting matrices allow one to emphasize the specific variables related to the vibrations of the specific bands of frequencies. Once the active control law is obtained, we use the artificial neural networks to train the neuro controller to learn the relation of road input and control force.

From the numerical results, we found that back propagation learning does good pattern matching and the neuro controlled suspension may reduce the vertical acceleration of the driver's seat and sprung mass motions significantly at desired bands of frequencies.