Research On Obstacle Avoidance Control Algorithm For Autonomous Vehicles

The main research content of this paper is the obstacle avoidance control algorithm for autonomous vehicle; it is mainly used to control the vehicle to avoid the static obstacles which can impact its normal driving in the road boundary. At present, most of the obstacle avoidance algorithm is based on the global information collected by system to make path planning and control vehicle to avoid obstacles. Thus, it is very necessary to study a kind of obstacle avoidance algorithm based on part information to control vehicles to avoid obstacles. The main research contents of this paper are as follows:Firstly, using monocular vision ranging method to locate the obstacle and road boundaries.The image collected by the camera through filtered, threshold, the edge detection processing to enhance the quality of the image, extracting the location of the road boundary in the image by hough algorithm, using the method of window retrieval to obtain the coordinates of obstacles in the image.The expression of the transformation between pixel coordinate and world coordinate is derived based on geometric relation, the camera is calibrated according to the parameters of the expression, complete parameter configuration, to achieve the localization of obstacles and road boundary.Secondly, according to the control parameters of the vehicle body and the features of application environment, a fuzzy control algorithm for obstacle avoidance is proposed. Take the horizontal and vertical distance between the vehicle with the obstacle, and the distance between vehicle body with road as the input of the control system, the steering gear output angle and motor speed of the vehicle as the output of system. The design process of fuzzy obstacle avoidance controller in this paper is discussed from three aspects, which are input fuzzy process, fuzzy rule establishment and defuzzification.Finally, by ranging experimental measure the horizontal and vertical distance of vehicle body with the obstacle, and the distance of vehicle body with the road boundary, the result of the measurement shows that the method of monocular vision ranging is effective and feasible in a certain extent. Then, through simulation experiment to verify the effectiveness of fuzzy control algorithm, simulated results show that the design of the fuzzy obstacle avoidance algorithm can make the car safely and effectively avoid obstacles in the road boundary.