Research On Path Planning Of Unmanned Ground Vehicles Based On GIS

Autonomous driving of unmanned ground vehicle(UGV) is a current hotspot of research on robot technology, and as one of its key technologies, path planning will ensure UGV drive fast, safely and accurately only by providing a feasible and optimal path between two waypoints.There is no doubt that it is difficult for UGV to perceive, since UGV has to perceive too much attribute information of the driving field in increasingly complex urban environment and unknown variable cross-country environment. As a particular spatial information system, geographic information system(GIS) with strong geographical information storage and analysis capabilities can effectively reduce the difficulty of environment perception.For UGV’s movement characteristics and various driving environment, based on GIS, this paper discussed path planning technology for UGV. The main research results are as follows:1. A GIS database was built and applied to UGV’s autonomous driving. It is used to store the priori information which can help the perception and decision of UGV, such as lane line information, traffic lights information, traffic signs, etc. And also it can provide the road map.2. In view of urban environment with structured road characteristics, the road matching method was firstly studied. Several matching methods were put forward, including the candidate sections matching algorithm based on computational geometry, task points matching algorithm and real-time waypoint matching strategy. Secondly, according to UGV’s movement characteristics, an improved double crossing model was put forward to strengthen the planned path trajectory practicability. Finally, a local path replanning strategy was developed which will help UGV get rid of the driving dilemma effectively and continue safe drive.3. In view of cross-road environment with unstructured road characteristics, a global path planning algorithm based on cubic Hermite interpolation was proposed by comparing different mathematical interpolation methods, which can make the movement track of UGV more pliable. Secondly, a local path replanning strategy was presented to ensure that UGV can pass the fork road smoothly. Finally, the methods of waypoints guidance and control were developed to ensure the convenience and easy operation of UGV’s driving combined with the actual application conditions.The validity of the algorithms which were proposed in this paper have been proved effective by a large number of actual car experiments. The unmanned driving car IN2 BOT based on these algorithms participated in “The Future Challenge: Intelligent Vehicles and Beyond” held by the National Natural Science Foundation of China and “Crossing Obstacles- 2014, First Challenge of Ground Unmanned Platform” held by the PLA General Armaments Department in 2014, and achieved good results.