Online Adaptive Full Car Active Suspension Control Using B-spline Fuzzy-Neural Network

In this paper, the Adaptive B-spline Fuzzy Neural Network (ABFNN) based an active suspension system for full car is presented. The passive suspension system cannot reduce the vibrations which are transmitted from the road disturbances to the frame which affect the ride comfort and vehicle stability. The magnitude of these vibrations can be reduced by using ABFNN based an active suspension system. The ABFNN has ability to approximate the nonlinearity of the vehicle. By using B-spline membership function in the fuzzy neural network the approximation ability of the network is increased. The shape of B-spline membership function is adjusted selfadaptively by changing control points during learning process. B-spline membership functions give a structure for choosing the shape of the fuzzy sets. The update parameters of ABFNN are trained by gradient-based technique that may fall into local minima during the learning process. The ABFNN is successfully applied to full car suspension model which reduces the seat, heave pitch and roll displacement of the vehicle. Simulation is based on the full car mathematical model by using MATLAB/SIMULINK. The simulation results show that the ABFNN control technique gives better results than passive and semi-active suspension systems.