Research On Road Recognition And Tracking Control Of Vision-Guided Intelligent Vehicle

Intelligent vehicle is a system integrated with the functions of environment-sensing, decision-making and automatically driving. It has been an important part of the Intelligent Transportation System and widely used in flexible manufacturing and resembling system, military field and special environments. The research on intelligent vehicle can improve the safety and comfort of vehicle driving, and the efficiency and environment of transportation.The key technology of the intelligent vehicle studied here includes control architecture, image recognition and road tracking. The control architecture is the framework of the intelligent vehicle for information-processing and control and the base for the whole system. Based on analyzing the characteristics of the existing architecture, an activity-based architecture for intelligent vehicle navigation is presented. In this architecture, the functions of the vehicle are achieved by performing one or more so-called activities, which make decisions parallel using distributed knowledge. And a blackboard coordinates the activities, makes final decisions and activates appropriate activities to control the vehicle so as to accomplish the navigation task.Based on the architecture, the vision-guide intelligent vehicle system based on road navigation is designed. The system is composed of the road recognition system and tracking control system. The road recognition system recognizes the road using vision sensor. Its key is to make sure the road is robustly and quickly recognized. So filtering, image segmenting using optimal threshold, morphological operation are performed to improve recognition accuracy, and some measures are taken to increase processing speed.The road tracking system controls the vehicle to track the road recognized before. Its key is the design of the controller. As the feedback controller uses only the present road information and can hardly work efficient under complicated or emergent conditions, the whole navigation road are classified into present road and future road, and they serve for feedback control and preview control respectively. So a feedback and preview controller is designed and it can adjust the preview distance and speed of the vehicle according to the road. As a result, the tracking accuracy and stability can be improved.Finally, the whole intelligent vehicle experiments are carried out in order to verify the presented theories and methods. The experiments under different conditions show that the designed vehicle system can recognize the road image accurately and quickly, and can efficiently achieve the goal of tracking different roads with different speeds.