Abstract
In this study, a vision-based driver-copilot system is developed to aid a driver to keep the vehicle; within the driving lane. In the proposed driver-copilot system, an image sensing technique is used to measure the motion variables of a vehicle that is, heading angle, lateral deviation and yaw rate. The road image features are used as the inputs and steering wheel angle is used as the output. An intelligent lateral controller is implemented using a table-lookup scheme for adaptive fuzzy system design and is used to perform a one-pass operation on the numerical input-output pairs and linguistic fuzzy IF-THEN rules. A fixed-base driving simulator composed of a real half-car steering mechanism is used for simulating the lane keeping control. By the vision-based estimation method, the driver can control the driving simulator through the steering wheel according to the road scene displayed in the monitor. The experimental results show that the developed vision-based intelligent driver-copilot is feasible.