Research On Cloud Platform Construction And Cloud Service Scheduling Algorithm Of Service Robot

Traditional service robots have two major technical bottlenecks in resources and intelligence,which are difficult to meet actual application requirements.To break the above-mentioned bottlenecks,it's an effective mean to build a service robot cloud service platform via combining cloud computing technology and robotics technology,and provide intelligent cloud services for robots.The service robot offloads ontology-intensive computing tasks to the cloud,and its intelligence level is expanded on a large scale and exceeds the limits of its physical ontology,providing active intelligent services for humans.The work of this paper comes from the national key research and development project,focusing on the specific research on the construction method of the service robot cloud service platform and the large-scale cloud service efficient scheduling algorithm.First of all,this paper constructs a general new type of basic service robot cloud platform,designs its gateway module,interface layer,service pool module and algorithm layer in detail,and proposes a standardized cloud service access and invocation method.Besides,the availability of the service robot cloud platform,the accuracy of the results,and the low latency of invocation are verified via empirical experiments,such as interface layer service monitoring and voice recognition service invocation.Secondly,this paper designs the related attribute models of service robots and cloud services,transform the balance process of the cost models of robots and service platforms into a single-objective optimization problem.In addition,we propose a hierarchical predation scheduling algorithm(RHGA),which is drawn on the law of prey-predation in nature,and adopt prey-predation hierarchical evolution,population evolution Factors and population update based on similarity are three strategies to improve the performance of evolutionary algorithms.The cost and time scheduling comparative experiments demonstrate that the superiority of the algorithm proposed in this paper in terms of service scheduling completion time and cost reduction.Finally,fully considering the characteristics of cloud platforms,service robots and cloud service developers,a cost tuple parameter model is constructed,and a dynamic vector hybrid genetic algorithm(DVHGA)combining local and global search processes and three-phase parameter update strategies is proposed.To ensure the reasonable allocation of resources,while optimizing the final cost of the two-stage service selection strategy constituting the model,and propose a linear evaluation method based on time and final cost.Experimental results show that the algorithm is better than other benchmark algorithms in terms of convergence speed,final total cost,evaluation score and algorithm simplicity.