Research & Implementation Of Path Planning Based On Feature And Generalized Topology For Motor Vehicle In City

With the expansion of city, requirement of the rapid urban transport is more urgent as well as the increasingly accelerating pace of people's lives. However, the new expansion of the road is limited by space, and it is impossible to satisfy social needs in an sizable city only relying on the construction of roads to solve the traffic problem. Thereby some advanced technologies which effectively use the resource have emerged to resolve road traffic problems. This paper mainly researched how to choose a more reasonable route in a complex city road network.The city path planning system was based on the Knowledge base management, Machine learning and Heuristic search in the field of Artificial Intelligence. It was developed by stretched rubber band algorithm which combined with A* algorithm and previous learn results. And this system used Borland C++ Builder as a development platform. It was oriented the motor vehicles and robots path planning in vast region. When people attempted to search for the optimal routing for motor vehicles or robots, they were all faced with how to avoid the impact made by multitudinous factors as the road network extensity, the complexity, the configuration uncertainty and time-variable. Under the premise of meeting users' different pursuits of performance, this system applied the most simple way to provide the action order which could practically guide acts, the action order was a integration of qualitative and quantitative, and had a good self-learning feature. Based on the selected benchmark of evaluation, users only needed to specify the start and goal, then system could automatically find out the best route.The paper first introduced the basic theory of artificial intelligence which this paper based on, and then it analysed the design of entire system, the frame, and the function modules carefully. The stretched rubber band algorithm could be used as a heuristic information of A* algorithm. The paper taked the city traffic as the experimentation background, it solved the issues of specific application, which used the three level searching at the key blocks passed, the blocks boundary points and the road network. In addition, the system used machine learning to improve performance. Later, the paper introduced how the city path planning system was designed and realized in detail, and the run results of the system was presented. Finally, the paper pointed out deficiencies in the system and the work in the future.