智能变电站四足巡检机器人狭窄通道路径规划OA

[Objective]The internal environment of intelligent substations is complex,with narrow channels varying in shapes and sizes.The complexity and uncertainty of these narrow channels require robots to frequently adjust their poses when they are passing through,thus increasing the path planning difficulty.Therefore,a narrow channel path planning method for quadruped inspection robots in intelligent substations was proposed.[Methods]The sensors inside the quadruped inspection robot were employed to estimate its own state,and external sensors were combined to achieve perception and positioning of the surrounding environment.By adopting the probabilistic positioning principle model,the pose estimation value of the robot in the intelligent substation was obtained by combining the observation results of LiDAR and mileage prediction results.Based on the estimated robot pose results,the problem of establishing a map was transformed into a maximum likelihood estimation problem of the map.Additionally,SLAM technology was utilized to process data,estimate the robot pose,construct a grid map,and update grid states.After completing the grip map establishment,the probability roadmap algorithm was adopted to plan the inspection path.The Gaussian sampler was employed as a roadmap sampling tool to randomly select a point in the grid map and collect a point along a random direction at a distance from the point.If the collection point is located in a blank grid,it is considered a collected roadmap point.Meanwhile,a Gaussian sampler was leveraged to collect a large number of sample points distributed around obstacles,with the number and contours of obstacles in the environment determined via the learning process.In this study,the starting and ending points were determined,and the number of sampling point nodes and the set of probabilistic roadmaps were set.The probabilistic roadmap set was initialized according to the starting and ending points,with new sampling points generated by sampling the grid map space.According to the probabilistic roadmap algorithm,a set of planned inspection path points were obtained,and the starting and ending points were set,with a set of path optimization points established.By employing the starting point of the path as the test point,the set of roadmap points on the path were tested one by one.If the test point cannot be connected to a certain point,it is considered a turning point,and is stored in the optimized point set and employed as a new test point.The roadmap points were tested backward one by one until reaching the ending point.After completing all tests,the roadmap points in the optimization point set were connected one by one to form a complete inspection planning path of quadruped inspection robots,thereby reducing the robot's pose direction conversion during inspection and ensuring the completion of inspection path planning without colliding with obstacles.[Results]The experimental results show that when this method is adopted for narrow channel path planning in substations,the number of sampling nodes is less than 52,and the average path cost is below 87 m.[Conclusions]The proposed method is validated to have sound planning effectiveness and superior performance.