2009 IEEE International Conference on
Systems, Man, and Cybernetics |
Abstract
The precise variation of the magneto-rhelogical (MR) damping force in a semi-active suspension is a key issue in order to assure the desired performances over a suspension control system. The open loop control of this force is a very common strategy. Other schemes propose adaptive controller alternatives while the automotive hardware is a constrained computation resource. This paper proposes the implementation of the damping force control system based on the MR damper using an internal model control approach. The controller and the internal model are proposed as artificial neural networks (ANN) trained and validated with realistic automotive datasets. The results shows good servo control and fast regulation to abrupt disturbances without on-line ANN tuning computations.