Research And Implementation On QoS-Awareness Software Platform Supporting Cloud Robotics

"Platform as a service" as one of the classic application of cloud computing model,has very important meanings for the rapid construction of cloud robot software and ecosystem technology.In the field of traditional cloud computing,the key technologies of "Platform as a Service" and the corresponding software infrastructures have accumulated a great deal.However,unlike traditional software system,robotic system is a typical system of deep integration of information and physics,which not only perceives but also influences the physical world.Therefore,many constraints of the physical world will be directly mapped into the robot software,showing a series of requirements for quality of service(QoS)such as real-time performance and request response time.This feature raises new requirements for the software platform that supports the operation of robotic cloud services to provide corresponding common support mechanisms for robot applications in terms of QoS.This is one of the major challenges that cloud robots currently facing with.In response to the above challenges,this thesis focuses on the key technologies of cloud-based robotic software platform that supports QoS,and then verifies it with the prototype system of software platform.The work of this paper includes the following three aspects:(1)Resource Guarantee Mechanism of Robust Cloud Service Based on Resource SchedulingRobotic cloud computing platforms need to be able to support concurrent access of a large number of clients and to provide QoS capabilities in this scenario.At the same time,robot applications have significant personalization features.Even for the same service and different application scenarios,different robot models often have different requirements on service quality.In order to solve this problem,this paper proposes a cloud service QoS guarantee mechanism that supports on-demand scheduling of resources.Through Servant's dynamic instantiation,resource allocation based on QoS dictionary,resource isolation based on lightweight VM container to guarantee the QoS of cloud service.(2)QoS Assurance Mechanism for "Cloud-Network-Robot" Collaborative Cloud Robotic Software SystemThe cloud robotic software system is a typical distributed software system,on which QoS guarantee needs the support of all aspects of distributed computing to form the corresponding end-to-end QoS assurance chain.In response to this problem and to achieve QoS Guarantee,this paper proposes a QoS-oriented mechanism of collaborating between robot and cloud through the automatic construction of cloud-service-accessing client(Stub),client-side compensation mechanism and the collaboration of client and cloud.In order to solve the problem of time optimization in the process of network message transmission,a framework of message transmission supporting compression has been designed further in this topic.(3)Prototype Realization and Experimental Verification of Cloud Robot Software Platform Supporting QoSOn the basis of the above mechanisms and based on Docker and Robot Operating System,this paper implements the cloud robotic software platform Cloudroid.The Cloudroid platform was tested for the transparent deployment and access capabilities and QoS assurance capabilities based on third-party typical applications.