| With the increase in the number of domestic tourists and the increasing number of bags checked in every day,the existing airport baggage handling system is struggling to meet the growing demand for baggage check-in.In 2021,the Civil Aviation Administration of China proposed that the development of "smart airport" construction will tend to be unmanned equipment,which makes autonomous mobile robots(AMRs)replace traditional conveyor belts to complete baggage check-in work a new format.Aiming at the problems of long path planning time and AMR congestion and collision in airport baggage tray scenarios,this thesis studies the multi-AMR path planning algorithm for airport baggage check-in,improves the path planning efficiency and solution quality by improving the Continuous-Time Conflict-Based Search(CCBS)algorithm in different directions,and verifies the efficiency of the improved algorithm through simulation experiments.The main research contents are as follows:(1)Airport baggage check-in scenario analysis and task design.A raster map of the airport baggage drop-off scene was constructed,and AMRs prone to conflicts in this scenario were analyzed.According to the different task types in the airport baggage check-in scenario,25 task files are designed,and each file contains 35 AMR tasks for the simulation experiment of airport baggage check-in.(2)Multi-AMR path planning search algorithm based on mutex.Aiming at the problems of incomplete cardinality collision recognition in CCBS algorithm,a multi-AMR path planning search algorithm based on mutex is proposed.By constructing a multi-valued decision graph under continuous time,the mutex propagation technique is used to identify the special cardinality conflicts between the two multi-valued decision graphs,which enhances the discrimination ability of cardinality conflicts.(3)Multi-AMR path planning search algorithm based on heuristic.Aiming at the problem that the geometric collision detection mechanism in the multi-AMR path planning search algorithm based on mutexes takes a lot of time and the high-level best search method selects nodes blindly,multi-AMR path planning search algorithm based on heuristic is proposed.The algorithm reduces the number of path searches by adding mutexes to higher-level functions and constructing heuristics with minimal vertex coverage.(4)Experimental verification of multi-AMR path planning simulation.The effectiveness of the proposed algorithm is verified by comparing the performance analysis of CCBS algorithm,multi-AMR path planning algorithm based on mutex,CCBS algorithm with least vertex overlay and multi-AMR path planning algorithm based on heuristics on blank maps,random obstacle maps and airport baggage check-in maps with different branching factors.Experiments show that the multi-AMR path planning algorithm based on heuristics performs best overall,and the number of high-level nodes is reduced by 99.7% and the average calculation time by 99.2% compared with the CCBS algorithm in the best case,and the multi-AMR path planning algorithm based on mutual exclusions performs best when the branching factor is 4 in the airport baggage check-in map,which is slightly better than the heuristic-based multi-AMR path planning search algorithm. |
