A Method Of Multi-handling Robot Scheduling In A Cloud-edge Environment

The rapid development of the express industry has put forward higher requirements for the degree of warehousing automation and efficiency.At present,some domestic e-commerce and logistics companies have gradually adopted multi-handling robots that integrate pick-and-place,storage,and transportation functions to replace shelf handling robots for warehouse automation transformation.The use of this kind of multi-handling robot is expected to further enhance the degree of warehouse automation and work efficiency.The task scheduling problem and path planning problem of multi-handling robots are NP-hard.To further improve the solving efficiency,this paper studies the parallel algorithms for solving multi-handling robots task scheduling and path planning,and designs and implements a network secured multi-robot scheduling software using a cloud-edge framework.The specific research content and work are listed as follows:Firstly,a parallel method for solving task scheduling of multi-handling robots is studied.The working mode of the multi-point handling robot is analyzed,and its task scheduling problem is modeled as a colored traveling salesman problem(CTSP).Based on the variable neighborhood search(VNS)algorithm,an improved VNS(IVNS)algorithm is proposed by combining the distance-based candidate set and the optimal 2-opt local search algorithm.Based on the thread pool,a parallel algorithm framework for solving CTSPs is proposed for the first time and a parallel IVNS(PIVNS)algorithm is designed using the framework.Besides,the IVNS algorithm can obtain a better solution compared with the three genetic algorithms(GAs)and VNS algorithm.Moreover,PIVNS can obtain better solutions and converge faster than IVNS.Secondly,path planning methods of multi-handling robots are studied.Single robot path planning methods are studied including the Dijkstra algorithm and A*algorithm.An improved A*algorithm is proposed by considering the steering of the robot.Using a single-robot path planning method to complete multi-robot path planning suffers from dimension explosion.Therefore,the multi-robot path planning method based on conflict-based search(CBS)algorithm is studied.Moreover,a parallel CBS(PCBS)algorithm is designed.The high layer of the CBS algorithm converts path conflicts into constraints,and the lower layer executes single-robot path planning under multiple constraints.By comparing the CBS algorithm using Dijkstra,A*and improved A*algorithm as the lower layer,it shows that the CBS algorithm with improved A*is the most efficient.In addition,the comparison of CBS and PCBS algorithms shows that the PCBS algorithm is more efficient.Thirdly,the requirements of scheduling software were analyzed,and the multi-handling robot scheduling software under the cloud-edge computing architecture was designed and implemented.The software is mainly divided into cloud task scheduling module,edge path planning module,and security protection module.Specifically,OpenFaaS is used to turn the PIVNS algorithm into a cloud service,and an edge controller based on the PCBS algorithm is implemented to complete multi-robot path planning.Simulation experiments of multi-handling robots verify the effectiveness of the scheduling software.The efficiency of cloud scheduling is verified by comparing the time-consuming of cloud and edge task schedulingFinally,the privacy leakage risk and cloud service interruption risk of the scheduling software using cloud-edge architecture are analyzed.Data encryption technology and transport layer security protocols are used to encrypt and protect the communication between the cloud and the edge,and an edge controller is used to complete data desensitization to prevent private information misuse.In addition,considering that distributed denial of service(DDoS)attacks can cause cloud service interruption,TCP configuration is modified to mitigate DDoS attacks.Then,to ensure the scheduling software availability under the cloud service outages situation,container migration technology is used to migrate cloud services to edge computing nodes when the service is interrupted.