A brake system controller is designed using the powerful techniques of neural networks and genetic algorithms. First, the problem of coordinating auxiliary brakes, foundation brakes, and gear for high transport effectiveness in down hill cruising situations is investigated. An optimization problem with constraints such as vehicle speed and disc temperature is formulated and solved, resulting in a well performing controller even compared to experienced drivers. Second, the issue of distributing a required force between auxiliary and foundation brakes in order to minimize the maintenance cost is investigated. The neural network controllers obtained from the optimization procedure significantly outperform the traditional strategy of using non-wear auxiliary brakes in order to minimize pad and disc wear cost. The performance of the brake system can be improved by controlling the whole brake system including gear. (A)
[1]
S. Hyakin,et al.
Neural Networks: A Comprehensive Foundation
,
1994
.
[2]
Bengt Schmidtbauer,et al.
Road Slope and Vehicle Mass Estimation Using Kalman Filtering
,
2002
.
[3]
X. Yao.
Evolving Artificial Neural Networks
,
1999
.
[4]
Reid Kress,et al.
A TOOL FOR DOWNHILL ACCIDENT ANALYSIS AND BRAKE DESIGN EVALUATION
,
2001
.
[5]
Anil V. Rao,et al.
Practical Methods for Optimal Control Using Nonlinear Programming
,
1987
.
[6]
Stephen P. Boyd,et al.
Convex Optimization
,
2004,
Algorithms and Theory of Computation Handbook.