Monitoring, modeling, prediction, and control of the braking process is a difficult task due to a complex interaction between the brake contact surfaces (disc pads and brake disc). It is caused by different influences of braking regimes and brake operation conditions on its performance. Faster and better control of the braking process is extremely important in order to provide harmonization of the generated braking torque with the tire-road adhesion conditions. It has significant influence on the stopping distance. The control of the braking process should be based on monitoring of the previous and current values of parameters that have influence on the brake performance. Primarily, it is related to the disc brake actuation pressure, the vehicle speed, and the brake interface temperature. The functional relationship between braking regimes and braking torque has to be established and continuously adapted according to the change of mentioned influencing factors.
In this paper dynamic neural networks have been used for the purpose of modeling and control of the disc brake actuation pressure. Parameters of the developed dynamic neural model were used to build a program for implementation in a microcontroller. Recurrent neural networks have been implemented in 8-bit CMOS microcontroller for control of the disc brake actuation pressure. Two different models have been developed and integrated into the microcontroller. The first model was used for modeling and prediction of the braking torque. Based on that, the second inverse neural model, has been developed able to predict the brake actuation pressure needed for achieving previously selected (desired) braking torque value.