Research On Task Planning Algorithm Of Autonomous Exploration Robot In Grottoes

Intelligent industrial robots will liberate people from tedious and mechanized work,but also improve production efficiency and product quality,has been successfully applied to various industries.As an important part,the exploration robot has been gradually applied in the cultural relic protection industry.Based on an intelligent robot developed by our research group for disease detection in grottoes,this paper mainly studies the task planning of autonomous detection function of the robot.The main work of this paper is as follows.Firstly,the autonomous detection function of intelligent detection robot in Grottoes Temple is analyzed.The autonomous exploration inside the grottoes temple is divided into two working conditions: elevation exploration and top exploration.The problem description of task planning is given,the algorithm requirements are put forward,and the system architecture of the two working conditions is designed.Secondly,the top surface detection task planning algorithm is designed,and its functions are divided into two tasks: image processing and task point extraction.The passable domain of the robot in the map is divided by image processing.On this basis,the robot activity space map is constructed,and the task point extraction is carried out on the passable domain in combination with the field parameters of the robot detection equipment and the requirements of full coverage detection.At the same time,according to the spatial structure characteristics of the grottoes and the safety requirements of the exploration operation,an improved simulated annealing algorithm was used to solve the task point sequencing problem.Then,the mission planning algorithm is designed,which is divided into two parts: the moving chassis task planning algorithm and the manipulator task planning algorithm.In the task planning of moving chassis,the sequence of moving chassis task points is extracted based on the idea of contour inward contraction and outward expansion.In the task planning of the manipulator,the elevation point cloud measurement scheme and point cloud processing and recognition algorithm are established,and then the surface feature direction recognition algorithm and region segmentation algorithm are designed.The point cloud slice algorithm is adopted for the dense area of point clouds with complex geometric structure and special treatment is carried out.Based on the principle of optimal observation and full coverage detection,The detection task points of the detection equipment at the end of the robotic arm were extracted.In addition,in view of the structural complexity of the internal facade of the grottoes temple,the K-Means algorithm was adopted to cluster the feature points,extract the fine detection task points from the clustering,reduce the rugged deformation of the images collected by the visual system,and further improve the precision of disease detection.Finally,the simulation and experiment of the task planning algorithm were carried out,and a ROS based physical simulation platform was built to verify the autonomous navigation ability of the robot based on the developed task planning algorithm,ensuring the feasibility of the application of the task planning algorithm to the exploration robot.Then,an experiment was carried out on the task planning algorithm of the intelligent exploration robot of the grottoes Temple in North Grottoes Temple in Gansu province,which verified the correctness and effectiveness of the proposed task planning method based on two-dimensional map and threedimensional point cloud from two aspects of the algorithm planning time and planning results.