Development of Lane keeping Controller Using Image processing

Abstract

Motivated by the need to overcome dynamic traffic congestion problems and driving safety issues, lane keeping control of vehicles has become a very active research area and is a driver-assistance vehicle feature that automatically controls a vehicle’s lateral direction at a certain look-ahead distance. This study presents the development of autonomous lane keeping controller using image processing in simulation environment and experimental setup. In the simulation study, a single tracked bicycle model was adopted as vehicle model. All simulation study of controller was based on the developed model in MATLAB environment while for experimental study; a scale vehicle was developed and used as a platform to test the lane keeping algorithm in a closed environmental setting. The lane keeping algorithm development comprises of two major parts. They are lane detection using image processing and lane keeping control. The lane detection provide important data like the lane boundaries and position of the vehicle with respect to the road and feed these data to the lane keeping controller for steering control. Proportional derivative controller (PD controller) control was developed to steer the vehicle and keep it on its lane at varying longitudinal speeds. Results in terms of stability of vehicle and maneuverability indicated that the range of values of both proportional and derivative gain altered the performance as can be seen in the simulation result and also suggested the inclusion of an Integral term I in an uncontrolled environment. The lane keeping algorithm was then implemented on the scaled vehicle and tested in a controlled environment and results indicated that the mobile robot was able to follow the desired reference trajectory by correcting the cross track error to the barest minimum. Experimental results also showed a similarity with simulations results of the mobile robot steering control