Research On Route Planning Methods For Intelligent Vehicles

Intelligent Vehicle (IV) is an advanced solution of the modern city traffic problems. Driverless IV can sense the environment around it and its own movement, then decide a safe, legal, and efficient track through a local area to drive the vehicle automatically. Anyhow, in order to reach some special destination in a large-scale map, avoid driving blindly, a navigation system is necessary.There're three important modules in a vehicle navigation system: digital map database, route planning, and route guidance. Above all, a digital map, also known as Navigable Geo-Spatial Database is needed. For an intelligent vehicle, large-scale route planning and accurate route guidance demand a road network topology structure, other than roads'shape and roadside environment. A Road Network Definition File (RNDF) used by IV challenges makes a good example. By analyzing the useful information of road network and traffic environment, consulting the standardization of Geographic Data File, combining with the format of RNDF, an expression of'Road Network Data with Integrate Information'formed by multi-levels is designed in this thesis.In the next place, route planning is a key module and essential problem in the vehicle navigation system. In substance, it is a graph searching problem. Although global route planning can be done once after a destination is given, a real-time dynamic route planning is still needed in case of where traffic information changes or accident happens. Many classic graph search algorithms are analyzed, tested, and contrasted in this thesis. According to the requirement of dynamic route planning, a basic graph searching algorithm is chosen for IV. Furthermore, an improvement is made on the algorithm in order to guarantee the accuracy and legality of the optimal route. To guide the IV more exactly without reducing the efficiency of the dynamic route planning, the planning method could be layered. The 1st layer only determines which segments of road or zones for the IV to pass through. The 2nd layer doesn't run until the vehicle approached some exact segments or zones, then outputs the exact route.The purpose of global route planning is to guide direction for the driverless IV, tracks in detail is decided afterwards according to the safety, legality, efficiency and so on. The route guidance also takes efficiency and legality into account. To guide the vehicle, an IV would pre-view the global route certain distance ahead, and evaluate the efficiency to approach the route. However, a car can't just run as fast as it can, so the IV has to judge the legality of the speed at the same time.At last, the efficiency and legality of the global route must be proved. By programming and testing the'Global Route Planning Method'based on A* algorithm, not only the optimal route is confirmed, but also the validity and integrality of'Road Network Data with Integrate Information'is approved.This thesis mainly provides a solution of navigation for driverless intelligent vehicles. In attempt of solving this problem, some innovations are made on some key technologies, such as road network database, route planning, and route guidance.