| In recent years,robotic automated storage systems have attracted a lot of attention.Currently,the traditional method of single-agent path planning is mostly used in robot path planning,by planning the shortest path,or minimum time path,for each robot individually.If two robots may collide during actual robot operation,the robots make their own judgments to yield or avoid,for example,one of the robots slows down or changes paths.This method works better in simple systems with a small number of robots.Once the number or density of robots increases,congestion or even deadlock can occur due to robots giving way to each other into traps,resulting in low efficiency of the whole system,which may even lead to production stoppage of the whole production line in serious cases,or even require manual intervention to solve the problem,affecting the overall system operation.The core work of this thesis includes the following aspects.(1)In the global planning stage,we compare the advantages and disadvantages of the jump point search algorithm and the ant colony algorithm,and choose the ant colony algorithm with long solution time but better solution results when the solution time is richer,and choose the jump point search algorithm with short solution time to obtain the near-optimal solution in emergency situations,so as to ensure that the response time of the system is shorter and the task route of the robot is relatively reasonable.(2)In the local planning stage,the vector field histogram(VFH~*)and the anticollision logic in the robot meeting can ensure no collision between robots and also deal with the unexpected situation such as the increase of obstacles due to the temporary change of the site map,so that the robot can actively make a detour in the sudden encounter of obstacles and actively give way or detour in the meeting,which can ensure the actual transportation of robots.efficiency.(3)By splitting the multi-intelligence cooperative operation into two stages,and using a combination of various unique path planning algorithms in global planning and local planning,the robot can avoid entering traps in local planning while ensuring the shortest running route,and can actively go around when encountering obstacles not found in the map,and can actively go around or give way when meeting traffic,so as to avoid congestion or giving way to each other to enter traps.Reduce the manual intervention,ensure the accuracy and efficiency of transportation,and improve the overall operation efficiency of the system.Through the research of this topic,the path planning of robot transportation in traditional robotic automated storage is improved,and the foundation for the subsequent implementation of multi-intelligent body path planning is established to provide an idea for the development of highly intelligent and fully autonomous intelligences in the future. |
