A neural network based anti-skid brake system

Pages: 321 - 338, Volume 5, Issue 2, April 1999

doi:10.3934/dcds.1999.5.321       Abstract        Full Text (514.7K)       Related Articles

Sanjay K. Mazumdar - Department of Electrical & Electronic Engineering, University of Adelaide, South Australia 5005, Australia (email)
Cheng-Chew Lim - Department of Electrical & Electronic Engineering, University of Adelaide, South Australia 5005, Australia (email)

Abstract: The novel application of a neural network based adaptive control scheme to an anti-skid brake system (ABS) is presented in this paper. The anti-skid brake system represents a unique and challenging application for neural network based control schemes. The principal benefit of using neural networks in anti-skid brake systems is their ability to adapt to changes in the environmental conditions without a significant degradation in performance. In the proposed approach, the controller neural network is designed to produce a braking torque which regulates the wheel slip for the vehicle-brake system to a prespecified level. An enhanced reference model is proposed which generates the desired slip response and enables a sufficient condition for the convergence of the tracking error to be derived. Simulation studies are performed to demonstrate the effectiveness of the proposed neural network based anti-skid brake system (NN-ABS) for various road surface conditions, inclines in the road, and transition between road surfaces.

Keywords:  Neural network, control scheme, anti-skid brake system.
Mathematics Subject Classification:  92B20, 93C10.

Received: October 1997;      Revised: December 1998;      Published: January 1999.