Development of nonlinear real-time intelligent controllers for anti-lock braking system (ABS)
Date
2013-02-13
Authors
John, Samuel
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The objective of the Anti-lock Braking System (ABS) is to control the wheel slip to maximize
the friction coefficient between the wheel and the road, irrespective of the road surface
and condition. The introduction of new braking system in road vehicles such as the electromechanical
brakes used in brake-by-wire (BBW) system, which has a more continuous
braking operation with a high level of accuracy, necessitates the continual review and improvement
of the anti-lock braking system. From the control view point, therefore, more
refinement of the ABS operation could be achieved with these improved hardware components.
This thesis proposes a hybrid controller; combining feedback linearization and
proportional, integral and derivative (PID) controllers, and a neural network-based feedback
linearisation wheel slip controller. Furthermore, the thesis investigated the viability of
a hybrid system of the proposed neural network and a (PID) wheel slip controller system.
The hybrid systems, combines the accuracy of slip tracking ability of the PID controller and
the robustness of the feedback linearization controller to achieve shorter stopping distance
and good slip tracking. The performance of the proposed ABS systems are validated in
software simulation and on a laboratory ABS test bench. The results for both controllers revealed
their robustness to different road conditions and good slip tracking. This work further
confirms the feasibility of a future neural network-based ABS controllers in road vehicles.
Keywords: Anti-lock braking systems, Wheel slip, Friction models, Neural networks, PID
controller, Feedback linearization controller, Intelligent controller, Hybrid controllers, Realtime
embedded systems.