Artificial Intelligent-Based Feedforward Optimized PID Wheel Slip Controller Samuel John School of Engineering The Polytechnic of Namibia, (Namibia University of Science and Technology) Windhoek, Namibia. Email: sjohn@polytechnic.edu.na Abstract—Continual improvement of the anti-lock braking system control strategy is the focus of this work. Advances in auto-electronics and sub-systems such as the brake-by-wire technology are the driving forces behind the improvement of the anti-lock braking system. The control strategy has shifted from speed-control to slip-control strategy. In the current slip-control approach, proportional-integral-derivative (PID) controller and its variants: P, PI and PD have been proposed in place of the bang- bang controller mostly used in commercial ABS. Though the PID controller is famous due to its wide applications in industry: irrespective of the nature of the process or system, it might lead to limited performance when applied to the ABS. In order to improve the performance of the PID controller, a neural network inverse model of the plant is used to optimize the reference input slip. The resultant neural network-based PID ABS is then tested in Matlab R  /Simulink R  simulation environment. The results of the proposed controller, exhibits more accurate slip tracking than the PID-slip controller. I. INTRODUCTION Road conditions in most developing countries where this research is being undertaken, are poor. This places high performance demand on vehicle dynamic systems such as the ABS, suspension system, traction control, to mention a few. Most vehicles used in these developing countries are mainly imported from Europe, United States of America (USA) and Asia, where the road infrastructures are relatively much better. This has motivated the need to investigate other alternative and more robust controller schemes to the current commercially available bang-bang controller. During an emergency braking, the driver relies on the dynamic safety systems to avert an accident. If on the other hand, an accident do occurs, the static safety systems are activated to minimise its severity. The effectiveness of the braking system depends on the tyre-road contact forces. These forces vary with respect to road condition, type of tyre and weather conditions. The role of the anti-lock braking system (ABS) as an active safety device, is to regulate the braking torque, in order to maximise the frictional forces between the tyres and the road. Consequently,the braking distance is minimise and the longitudinal stability and driver’s control of the steering is improved. This in effect avert skidding and enhances the driver’s ability to avoid obstacles. Most commercially available ABS controllers are table and relay feedback based. They are designed to work with hydraulic braking systems. These early systems operated on analog computers with vacuum-actuated modulators [1]. The current rule based controllers typically have several hundred rules that capture all braking manoeuvring [2]. The controllers, are tuned in a trial-and-error method, using simulations and several field testing. With advances in the auto-electronics and braking sub-systems: like the brake-by-wire; which utilises electromechanical braking mechanism, there is an opportunity for up-grading current ABS control schemes. In a brake-by- wire system, the actuators supply a more continuous and more accurate braking pressure to the four wheels independently. This system gives the advantage of controlling tyre slip at arbitrary set points which can be used to improve the control of the vehicle. Recent research works are based on slip-control rather than the rate-of-deceleration used in commercial ABS. In this new approach, the role of the controller is to track a desired set- slip value irrespective of changes in road conditions and other unexpected vehicle dynamics. Some control modes that are proposed in the literature, include the proportional-integral- derivative (PID) controller [3], [4], [5]: this choice of controller is obvious, as the PID and its variants: P, PI and PD are extensively used in industry [6]. The PID is applied to most processes whether linear or non-linear, due to the fact that the controller could be tuned by means of various available tuning methods, some of which are achieved by a means of a push of a button [7]. Tanelli et al [4] proposed a ‘nonlinear output feedback Proportional Integral control law’ for the ABS, the slip and wheel speed measurements are used as control inputs. The controller, according to Tanelli et al [4], does not require any knowledge of the current road condition or of the instantaneous value of the normal force exerted on the tyre. The work is based on the quarter-car model, and simulations are conducted on dry and wet asphalt road conditions. The results for different vertical forces reveal good performance of the controller. However, this work does not consider the suspension dy- namics. The suspension system plays a very crucial role in the dynamic changes of the vertical forces exerted on the tyres during braking in rough road terrains [2]. In another work, Fu et al [5] investigated the performance of ABS using PID controller on different complex road conditions. In this work complex surface is define as any road with alternate adhesion coefficient. Simulations were conducted on a surface that transits from low to high to low adhesion coefficient. With an initial speed of 50km/h, the PID controller recorded a