The Research On Path Planning Method Under Uncertain Environment For Mobile Robots

In recent years,as mobile robots are more and more used in unknown，complicated and adverse circumstances to help people to accomplish the workhuman beings cannot complete, the autonomy and intellectuality of moblie robotsused in unknown environments should be improved and enhanced. Independent pathplanning technology of mobile robots used in unknown environments is a basic anddifficult problem of realizing the autonomy and intellectuality of moblie robots.Thisdissertation focuses on the methods of independent path planning technology ofmobile robots used in unknown environments.The essential parts of independent path planning system are analyzed anddesigned pertinently. On the part of environmental information acquisition, asystematic comparative study of information acquisition methods based onstereoscopic vision,laser ranging sensor and3D camera is made,summarize everymethod’s advantages and disadvantages, and the methods based on3D camera andlaser ranging sensor are selected, the algorithm of describe terrain attributeinformation with the help of3D camera is proposed; On the part of location,research summary of locating methods is surveyed, finally the location method basedon inertial sensor is selected and designed; In order to analyze the motion trail ofmobile robots, movement model of mobile robots is built; And then, the systemarchitecture,basic function,work process of the path planning system’s softwareinterface are introduced.According to the diversity and position-uncertainty of obstacles in unknownenvironments, the mothods of remote path planning based on laser ranging sensorand close-range path planning based on3D camera are proposed. Laser rangingsensor is used to find out the temporarily target point, which would aviod the mobilerobots move to the long-distance impassable areas.The close-range path planning is based on the terrain attribute information gotby3D camera, this is also the main research contents of this thesis. The method ofsearching for sub-goal based on constraint points is proposed. Firstly, the obstaclesis classified and modeled, the types of obstacles are rocks, slopes and ditches. thepoints on obstacles which may be midway points of the planning path are summed upand classified into crossing points and bypassing points. With the help of A* algorithm, the path is planned between the start point and every constranit points,and then compute to get the optimal midway point. When mobile robot moves to themidway point, it would execute the crossing strategy or the bypassing strategyaccording to the classification of this constranit point. In this thesis, three crossingand bypassing strategies of three types of obstacles are designed. Besides, Back-to-original-point strategy is proposed in this thesis, when mobile robot moves into thearea which couldn’t find a path in front of it, it will move back to the original pointto finf another path. Simulation experiments are done to prove that the methodproposed in this thesis is effective and practical.