Research On Data Consistency And Task Scheduling Methods Of Service Robot Cloud Service Platform

With the increasing maturity of big data,cloud computing and artificial intelligence technology,it provides the foundation and conditions for the research and development of cloud-based intelligent robots.The service robot cloud service platform(smart cloud brain)is the core and foundation of cloud-based robots,which is still in the initial research and development stage,and there are problems that need to be solved urgently,such as poor realtime system response and high service scheduling complexity.In view of the above problems,this paper analyzes the requirements,design architecture and functions of cloud service platforms,and conducts in-depth research on the data consistency methods and efficient service scheduling algorithms of cloud service platforms.The main work carried out in this paper is as follows:In order to meet the access requirements of large-scale heterogeneous cloud-based service.robots,a general cloud service platform system architecture based on cloud-edge-end collaboration is designed,and four functional layers of the system are designed in detail according to the functional requirements:the cloud service infrastructure layer responsible for computing and storage;the cloud service layer responsible for the construction,storage and invocation of cloud services;the API gateway layer responsible for the unified entrance of the platform and the user interaction layer responsible for managing platform users.The test results verify the effectiveness and availability of the cloud service platform architecture.Aiming at the problems of large consistency delay in cloud data transmission and unbalanced network connection status of clients in local area networks,a data consistency method based on cloud-edge-end collaboration is designed.Firstly,the grouping technology is used to adaptively determine the best quorum set of the cluster,which reduces the number of nodes of the two-stage quorum set and reduces the data consensus delay under the premise of meeting the fault tolerance requirements of the system.Then,an adaptive leader selection model is built to actively transfer the leader node to the region with high request frequency,so as to reduce the communication delay between the leader node and the client.The experimental results show that the communication delay of the proposed method in cloud platform data transmission is reduced by 11.43%compared with the best comparison method.Aiming at the problem of high scheduling complexity of cloud services under high concurrency,a two-stage rate monotonic algorithm with improved priority is proposed.Firstly,a multi-impact factor analytic analytic model is designed to obtain the task synthesis priority.Then,two task sets are classified according to priority,which improves the schedulable range of monotonic rate algorithm.Finally,the two-stage monotonic rate algorithm is adopted to solve the problem of cloud service scheduling under limited resource constraints.The experimental results show that compared with the current optimal comparison method,the proposed algorithm can reduce the completion time and loss by 9.55%and 11.29%,respectively,and the value and reliability can be improved by 4.85%and 2.58%,and the task preemption of the monotonic rate scheduling algorithm is reduced by 45%when the number of tasks is equal to 100.In order to verify the running effect of the cloud service encapsulation call model,data consistency,and service scheduling algorithm proposed in this paper in the production environment,the algorithm model is deployed as a standard component to the public cloud environment for instance operation verification.The application results show that in scenarios where the cloud service is called 1-200 times,the success rate of the cloud service response is 100%,and the response period is less than 1 second.In the real-time push scenario of video stream data,the communication delay decreased by 13.25%,and all requests were successfully responded.When the number of nodes is 10,30,and 50,the average response time of highconcurrency static service requests is reduced by 0.167 seconds,0.187 seconds,and 0.213 seconds compared with the original optimal scheduling method.