Research On Path Optimization Of Suspended Track For Multicopter UAV Inspection

Suspension monorail is a new type of urban rail transit system,with independent right of way,short construction cycle,low cost and operating cost,safe and comfortable,green energy-saving and non-pollution characteristics,and has vast development foreground in China.Nevertheless,the working environment of the track is mostly outdoor.Long-term exposure to the local atmospheric environment has made the protective coating of the track paint easily damaged,and the probability of track wear,rust and even broken greatly increases.All kinds of electronic equipments installed on the track also have the risk of failure.In order to protect the personal safety of passengers and the normal operation of the system,it is very necessary to carry out regular inspection of the track.One remarkable feature of the track lies in its suspension in the air.Thus,if it adopts manual method to troubleshoot the track,there will be disadvantages such as long inspection cycle,high cost,many blind spots,and low safety.In response to the above issues,this paper proposes a method of using multicopter unmanned aerial vehicle(UAV)to inspect the suspended track automatically.The specific contents are as follows:(1)A single-UAV inspection suspended track path optimization scheme is designed.Firstly,a mathematical model of the single-UAV inspection path optimization is constructed,which aims to minimize the UAV inspection path distance.On this basis,a brain storm optimization algorithm based upon variable neighborhood search(VNS-BSO)is devised to resolve the shortest inspection path.The algorithm introduces three neighborhood operations: swap,reverse and insert,to locally optimize the shortest inspection path obtained by the BSO algorithm at each iteration,and adopts an elitist preservation strategy to improve the search speed of the algorithm.In order to verify the feasibility of the scheme,twelve instances are selected as monitoring point data for 30 sets of simulation tests,and the best value,worst value,average value,standard deviation,relative error and convergence of the objective function are used as evaluation indicators.The results show that the shortest inspection path solution solved by the proposed algorithm has better quality,smaller dispersion,and faster convergence speed of the algorithm.(2)Aiming at the disadvantages of poor endurance,small monitoring range and easy environmental impact during the inspection of a single-UAV,a path optimization scheme of suspended track using multi-UAV collaborative inspection is proposed.Firstly,a mathematical model of multi-UAV inspection path optimization is constructed,in order to minimize the maximum inspection distance in all operating UAVs.Secondly,the multi-UAV inspection path is encoded and decoded,and on this basis,a large neighborhood search grey wolf optimization algorithm(LNS-GWO)is proposed to resolve the optimal solution of the multi-UAV inspection path.The algorithm introduces the destroy and repair operators in LNS,through locally optimizes the multi-UAV inspection path corresponding to the leader individual in each iteration of GWO,the whole path population evolves towards a better objective function value,so as to improve the accuracy and convergence speed of GWO search for the optimal solution.In order to verify the advantages of multi-UAV inspection and the proposed LNS-GWO optimization performance,four monitoring point location distribution scenarios are set.Firstly,the maximum inspection distance of the operation UAV is compared by the inspection efficiency comparison experiment when UAVs of different orders of magnitude perform these inspection scenarios,and the results show that within a certain number of ranges,the more UAVs participating in the inspection task,the shorter the corresponding maximum inspection distance and the higher the inspection efficiency.Secondly,20 sets of simulation experiments are carried out on the selected four scenarios using the best value,worst value,average value and standard deviation as the judging factors.The results show that the multi-UAV inspection path solution solved by LNS-GWO has better quality and less dispersion.The final convergence comparison experiment shows that the proposed LNS-GWO algorithm converges faster and has better convergence than GWO.(3)The latitude and longitude data of 20 suspended permanent magnet maglev rail transit system("Redrail")track monitoring points are collected through the designed monitoring point pickup interface,and the corresponding optimal inspection routes are planned by using the proposed single-UAV and multi-UAV inspection path optimization schemes,and the UAV is controlled by the ground station for actual flight testing.The test results show that the absolute error between the actual flight distance of the UAV according to the planned inspection path and the calculation result of the scheme is 4.87%,3.70% and 3.84%,respectively.The errors are all within the allowable error(5%),which verifies the effectiveness of the designed single-UAV and multi-UAV inspection path optimization scheme.