Abstract
Lane keeping is an important factor for driving safety and has become an essential issue of Intelligent Transportation System (ITS).This thesis designs a vision-based driver-copilot system for aiding driverto keep the vehicle within the driving lane. In the proposed drive-copilotsystem, the image features are used as the inputs and steering wheel angleis used as the output. Two intelligent lateral controllers are implementedusing the Adaptive Neuro-Fuzzy Inference System (ANFIS) method and thetable-lookup scheme. The learning algorithms are implemented tosimulate the human driving processes which is performed by an animation simulation. The simulated results show that it is feasible of using the vision-basedcontroller to keep the vehicle within a lane and the switching strategycan assist the driver in case that the vehicle is deviated away from thecurrent driving lane.