Research On Obstacle Avoidance Smart Car Sytem Based On Machine Vision

The study on obstacle avoidance of smart car has attracted more and more attention gradually. In order to improve the ability of the smart car’s barricade recognition, the machine vision has been applied to the road extraction in the barricade recognition process of the obstacle avoidance smart car, which promotes the development and application of the smart car.In this paper, the smart car system has been further explored and researched. Image capture technique was used to transform road condition into image signal, the barricade recognition and avoidance technique was integrated with it simultaneously to complete the smart car obstacle avoidance. Research on obstacle avoidance of smart car could provide ideas for the future realization of driverless car obstacle avoidance technique.This paper proceeded from the application of machine vision in smart car, designed the mechanical structure and modular circuit for obstacle avoidance smart car according to the specific requirements. With the MC9S12XS128as system control processor, a detail design for hardware circuit composed of power supply module, servo drive module, motor drive module, CCD camera detection module and the HMI module was given. As for software part, based on the road image information acquired by the camera image sampling module, using edge detection method to extract the road guide line, adopting dynamic template matching to extract the obstacle, the position between car and guide line, location between car and obstacle could be recognized when the car is running. According to the positional deviation between guide line and car, and the distance between the obstacle and car, an obstacle avoidance algorithm based on expert reasoning was proposed to determine the set points of the motor velocity and the servo direction. Then, in order to realize the obstacle avoidance during the smart car running, PID and fuzzy control technique are used to efficiently control the velocity of motor and the direction of servo, respectively. The simulation and test results show that the designed smart car can meet the requirements. Figures61), tables3), references(64).