Design And Implementation Of Service Robot System Based On Cloud Computing

With the development of the application field of service robots,service robots are faced with a lot of pressure of data storage and calculation when performing tasks.However,due to the limitations of hardware level,energy consumption and manufacturing cost,the local resources and computing power of service robots are often limited,which seriously restricts the autonomous behavior and service ability of service robots.Cloud computing has the characteristics of strong computing power,large amount of data storage and providing resources on demand.Therefore,combined with cloud computing technology,this thesis proposes a service robot system architecture based on robot operation system(ROS),which migrate real-time and compute-intensive tasks in the service robot to the cloud to reduce the consumption of local resource and reduce the cost of hardware.This thesis mainly studies the following points.First of all,this thesis analyzes the requirements of the service robot system,aiming at the migration of computing-intensive tasks in the ROS to the cloud,gives the overall design of the service robot system combined with cloud computing,and divides the system into service robot side and cloud platform,The perception,processing and execution of system structure and data interface are designed in the service robot side,and the system in the cloud platform side is divided into basic layer,platform layer and service layer.Secondly,this thesis focused on the communication of the system,collaborative computing between the service robot and the cloud,and the management of resource.In view of the fact that the message communication mechanism of ROS can not be directly transmitted through the Internet,this thesis realizes message conversion based on Rosbridge,which enables the service machine man to communicate with the cloud using Rosbridge and deliver ROS messages.For computing intensive tasks,it is simultaneous Location and mapping(SLAM)consumes too much local resources of service robots This thesis proposes a distributed slam service architecture based on cloud computing.By deploying ROS in the cloud platform layer as the application basis,the implementation of slam services is migrated to the cloud to solve the problem of local resource constraints.In the system resource management of service robots,the container based technology resources are proposed The management framework of service robot is designed to support multi service robot.Finally,this thesis implements the corresponding prototype system,and verifies the proposed architecture through experiments.The results show that the system can improve the computing performance of the service robot,save local resources,and meet the requirements of service robot management.