| The Repairing Day is a regular vehicle technical condition management system in the army.In the process of organization and implementation,the main tasks are vehicle maintenance,vehicle knowledge learning,etc.At present,the main problems are:improper methods,low completion quality and insufficient attention of personnel.In order to comply with the intelligent construction of the army,timely check the completion and timeliness of the system,and improve the organizational efficiency,the daily inspection robot of the parking lot is introduced for research.Its main functions are:automatic patrol inspection of the whole depot,inspection and counting of combat vehicles,supervising the operation soldiers to correct illegal operations,guiding the transmission of assistance information,regional material transportation and call response.The robot is required to travel safely and efficiently,with low energy consumption and good materiality and reliability.The path planning algorithm is the key and foundation to complete the above functions,which requires that the path planned by the algorithm has short length,high smoothness,maintains a safe distance from obstacles,and can avoid temporary obstacles.The parking lot environment has the characteristics of relatively stable basic layout,regularity,squareness and wide space in the middle.In this paper,RRT,genetic algorithm and A* algorithm are selected for planning in the parking lot simulation map.Through comparison and analysis,it is concluded that the A* algorithm has the best performance and is more suitable for the parking lot environment.Therefore,it is used as the basis of algorithm research.The specific research work is as follows:(1)Aiming at the problem that the traditional A* algorithm has poor performance and can not avoid temporary obstacles,expand the search neighborhood and remove redundant sub nodes in the same direction to improve the search speed and path smoothness;The risk index used to measure the safety of robot is introduced,the evaluation function of A*algorithm is redefined,the selection of nodes is optimized,and the distance from obstacles is guaranteed to form an improved A* algorithm;It is integrated with DWA algorithm to give dynamic obstacle avoidance ability.Finally,a hybrid algorithm of improved A*algorithm and DWA is proposed.Through the comparative experiment of the proposed algorithm,the original algorithm and the improved algorithm in the literature,the effect is analyzed.(2)Aiming at the problem that mapping algorithms affect the performance of path planning,this paper analyzes and compares different mapping algorithms such as Gmapping,Cartographer,Hector and Karto,builds a simulated parking lot environment in gazebo,compares and verifies the mapping,obtains the optimal mapping algorithm in the parking lot environment,and further proposes A* fusion DWA path planning algorithm under Cartographer mapping.Finally,the physical environment is built according to the physical characteristics of the parking lot,and the proposed algorithm is verified and analyzed.(3)Aiming at the problems of path conflict and planning failure in the parallel operation of multiple inspection robots.The proposed fusion algorithm under cartographer mapping is used as the single machine planning algorithm;Referring to the international maritime collision avoidance rules(COLREGs),a conflict resolution method is defined to form a multi robot path planning algorithm.Under the experimental platform of MATLAB and cartographer,the path planning algorithm without multi machine cooperation is compared with the encounter situation.The parallel operation experiment of multi depot daily inspection robot is carried out,and the typical complex situations are selected for analysis to verify the effectiveness and feasibility of the algorithm.Based on the above,a highly substantive and feasible path planning scheme is formed for the parking lot daily inspection robot to meet the needs of improving the efficiency of the parking lot daily system. |
