The Air-ground Coordination Path Planning Algorithm Research Via Visual Perception

In recent years,with the rapid development in the field of intelligent robots,a wide variety of intelligent robots rely on their unique skills to jump into the production of our daily life.Intelligent robots are playing an increasingly significant role in all walks of life while bringing great convenience and liberation to human.Especially in electric power system,many inspection and rescue robots can replace laborers to work in the high voltage and radiation environment.As a vital part of the intelligent robots,it is no doubt that the navigation and localization of mobile robots is valuable.Moreover,studying path planning algorithm of mobile robots can ensure robots to judge independently whether there exists the optimal path from start to finish.In this paper,we propose an approach to air-ground coordinated path planning based on visual perception.It is divided into three sub problems: the problem of local path planning with variable planning scale,the problem of path planning in three-dimensional space and avoid dynamic obstacles.The main work includes the following three aspects(9):(1)According to the designed scope,it is common for mobile robots to be divided into local path planning and global path planning,and the former associated with real-time is important to robot navigation under the unknown circumstance.However,most of the existing works just be carried out within a fixed size area,which is bound to cause a waste of time and resources in some cases.Therefore,this paper proposes a regional scale variable local path planning algorithm.The simulation results show that the proposed algorithm enables robots to determine the size of the designed area in accordance with the number of obstacles in the area,using a more intelligent way to plan a path.In addition,the proposed algorithm has better application in traffic road map.(2)At present,along with the rapid development of mobile robot industry,it is obvious that researches in the context of two-dimensional plane cannot satisfy planning requirements for some specific types of robots such as unmanned aerial vehicle and underwater robots.Therefore,this paper presents a novel mobile robot path planning algorithm in the three-dimensional space,provides a possibility to find a shorter path.In addition,previous researches are conducted in a static environment which is inapplicable in the practical condition that mobile robots encounter dynamic obstacles frequently.Therefore,this paper presents a real-time local path planning algorithm for mobile robots under dynamic obstacle environment.The simulation results show that the proposed algorithm making mobile robots have the ability of planning a path while avoiding dynamic obstacles.(3)In the unknown environment,the ground robots subject to the limit of their ability to sense surrounding.In some cases,the ground robots will be surrounded by obstacles without awareness,wasting more time and distance to get rid of trouble and get to the destination.Therefore,in order to avoid this kind of situation,this paper adds unmanned aerial vehicles(UAV)to provide high field of vision for ground robots,UAV first scan the unknown region in which ground robots will drive and inform them of obstacle information before starting.The simulation results show that the greater region UAV can scan and the more information they can provide,the better path planning effect ground robots can achieve.