Research On Hybrid Path Planning Of Substation Inspection Robot

Substation is an essential component of the power system,and its security and stable operation of the entire power system plays an important role.In order to assure the correct operation of the substation,the employees will regularly examine the devices in the substation,but the method of manual inspection has the drawback of large workload,inefficiency and misjudgment,which can not meet the needs of inspection work of modern substation.In addition,to realize the automation,intelligence and efficiency of inspection process,the inspection robot is brought into the electrical substation to support or replace the manual inspection.The path planning,an essential technology for inspection robots,also has been rapidly advancing in recent years.For ensuring that the inspection robot can trace the inspection points and shortest paths while running and escaping the detected moving obstacles,a hybrid path planning method that mixes global and local path planning is discussed in this paper.The major tasks are as follows:This paper first analyzes the inspection task of substation and working environment of inspection robot,and determines the positioning of robot by laser guidance on this basis.Simplified the substation environment and built its actual environment map model by grid map to offer environment data for path planning.The next,the improved ant colony algorithm is used for global path planning.In this article,we introduce a dynamic stochastic way to the transfer probability adjustment,and adding the direction factor of the artificial potential field into the information heuristic factor,which makes the algorithm can faster converge in the early stage of the search and can prevent from falling into the local optimum in the later stage.Replace one-way search with twoway search to enhance the efficiency of path search.Third,for the dynamics of the environment during the inspection,the improved Dynamic Window Approach(DWA)is employed for local path planning.The introduction of target distance into the evaluation function in this paper will allow the robot to increase the speed control capability and improved deviation from the target point when it is near the target point;combining DWA with the dynamic window method,LIDAR is used to obtain environmental information,to update the environmental map and to predict and analyze the movement trajectory of obstacle in real time within the rolling window,and take corresponding local path planning steps according to different collision types.Eventually,combining global and local path planning,we obtain the global optimal path by improved ant colony algorithm,conduct local path planning for dynamic obstacles encountered by improved DWA in traveling along the optimal path,and after completing obstacle avoidance,finally return to the global optimal path and travel toward the target point.This paper simulates the traditional ant colony algorithm,improved ant colony algorithm,improved DWA and hybrid path planning algorithm on MATLAB software,and performs the hybrid algorithm on Apollo experimental platform for testing.It is found that the improved ant colony algorithm is improved in convergence speed and optimal path length.Improve DWA can solve the question of easy deviation from the target point when close to the target point,and it predicts and analyzes the movement trajectory of obstacles and adopts corresponding obstacle avoidance measures based on different collision types.Mixed algorithms enable robots to quickly search for global optimal paths in complex environments and avoid dynamic obstacles encountered while driving.By using the hybrid algorithm,it is possible for the robot to perform path planning in the presence of unknown obstacles in the substation,to traverse the inspection points,to travel by the shortest path and to avoid dynamic obstacles.