Research On Cloud Robotic-based Simultaneous Localization And Mapping Technology

With the development of robot technology,the field of robot applications has been deepening.people demand for and the task of the robot is increasingly complicated and dense.At the same time,because of the limitation of size,power consumption,cost and so on,computing resources of the robot(such as the computing power of onboard computer)is often limited,which is unable to meet the needs of dealing with complex tasks.The rise of cloud computing technology has brought opportunities to solve this contradiction.Cloud Robotics has injected new potential for robot technology development,through in the form of "resources as a service" in the background support robot task efficiency.The combination of cloud computing and the robot has gradually become the research focus in the academia.Simultaneous Localization and mapping is a classical problem in the field of robot.That pays attention to how robot simultaneously perception environment map and determine its position in real time is considered to be the first step on the robot achieving real democracy.Traditional solution completely executed locally in robot.Due to the robot the limitations of its own resources,its execution speed and precision is difficult to meet the needs of actual robot application.This paper surrounding the solution of robot SLAM in Cloud,with which can make up for the inadequacy of robot local resources.In this paper,the main work embodies the following four aspects:1)Propose the SLAM service architecture of the Cloud Robotics.The architecture will put the processes with computationally intensive of SLAM such as data processing and data calculation into the cloud,which provided for one or more robot in the form of cloud services.Robot active application SLAM service at run time,continue to provide perception data to the service environment,and then get the result of building and mapping and its position without care of how SLAM process is to implement,weather computing resources are enough and so on.2)Design SLAM service prototype system of Cloud Robotics.In the guide of the SLAM service architecture of Cloud Robotics,SLAM service prototype system was implemented.The system can be divided into two levels,one is a “local + remote” collaborative SLAM task,is a simple model of robot request SLAM services to cloud;another is to realize the local active application service and the extension of the remote to form a cloud based on the model above,which has finished a complete implementation of the architecture.3)Design the optimization mechanism of data communication between the cloud and the robot.Under the architecture of Cloud Robotics,the robot continue providing RGB-D video data to the cloud,which is possible to exceed the bandwidth restrictions of robot wireless network.In this paper,we realize the optimization mechanism of data communication based on data compression in SLAM service prototype system of Cloud Robotics.4)Deploy SLAM service prototype system of Cloud Robotics over the ROS,which is a widely accepted robot operating system in the current.Also,we test for its functionality and performance evaluation based on the real robot and verify the benefits of the Cloud Robotics structure.