The Anti-Lock Braking System (ABS) is a safety critical feature primarily used to control slipping of wheels, to maximize available traction and minimize stopping distance. Regulatory authorities of many countries have mandated implementation of an ABS as a compulsory safety feature to be present in all road legal automobiles. Hence, apart from avoiding wheel lock-up, an ABS must also ensure that the vehicle maintains its handling stability and steerability while braking. Thus, it is important that the ABS controller modulate and apply adequate amount of brake cylinder pressure. This paper suggests the use of a Tire Force based algorithm to analyze vehicle behavior and accordingly a control law is employed to modulate the wheel brake pressure. A comparison study has been performed among control methods such as Step Gain Reduced Order Model (SGROM) developed using Machine Learning techniques and Linear Quadratic Regulator (LQR) to determine an optimal control law for brake pressure modulation under pure and combined slip condition.
The SGROM control technique implements a wheel cylinder pressure gain that is a function of vehicle state variables such as input brake pressure, normalized tire force and vehicle speed. The optimum output pressure gain is estimated by using a machine learning technique that is further elaborated in the paper. The LQR control technique regulates wheel cylinder pressure gains by minimizing cost function based on desired braking characteristics.
The control algorithms are evaluated based on standard brake performance tests such as stopping distance, split mu, and brake in turn. Accordingly, an optimal control strategy is selected based on evaluation criteria defined in the standards.