Research On Mobile Robot Autonomous Control System Based On Cloud Platform

Mobile robot is a intelligent system which combined many advanced technologies has been used widely such as server and rescue.With the development of cloud computing,it's a feasible solution that we can put large scale of data processing into cloud to solve the scarcity of mobile robots' local computing resources.Google's Waymo platform and Baidu's Apollo platform all take the use of cloud computing technology to process the data generated by the unmanned vehicle.This paper draws on their ideas and the cloud platform for omnidirectional mobile robot is designed which is based on them.We have analyzed exist omnidirectional mobile robot design cases which have a combination of cloud platform,the general framework of three wheeled omnidirectional mobile robot based on cloud platform and its autonomous control system are put forward.This paper has given the system model of omnidirectional mobile robot,finished the double closed loop attitude control on omnidirectional mobile robot,build cloud movement terminal platform which based on android system and Java Web,finished cloud computing platform which based on Spark and data exchange experiment between omnidirectional mobile robot and ROS center,take use of deep camera and Spark cloud computing platform to realize the autonomous control of omnidirectional mobile robot without GPS.Robots are a highly intelligent system that integrates multiple disciplines such as electronic technology,control science,communications engineering,mechanical design and manufacturing.The traditional mobile robot monitoring system is limited by the communication technology and the computer technology,the function is single and the use situation is limited.This paper constructs the mobile robot monitoring system based on the cloud platform,the overall architecture of the mobile robot monitoring system based on 4G wireless network and the distributed data storage platform based on Hadoop is designed.Real-time monitoring,real-time positioning,parameter tuning and other core functions and to achieve the mobile robot monitoring system data in the cloud platform on the distributed storage.From the hardware level,the mobile robot includes the car body,the motor and the bottom control unit,and is equipped with a depth camera,GPS/ Beidou and attitude sensors to detect the external environment.As the middle layer of the robot,ARM take the charge of the communication and human-computer interaction of the master control unit and the robot ROS host.From the software level,the omnidirectional mobile robot uses the ROS(robot operating system),the Apache Spark real-time computing engine,the Android development technology and the Web development technology to calculate and control the mobile robot.ROS as the core,mainly completes the information management of sensor information and motion control system and cloud computing of omni directional mobile robot.Android,Web and Spark communicate with ROS through ROS-bridge tools.Apache Spark subscribes to ROS packaged robot data and combines local programming to process the robot data and returns the result from ROS to the mobile robot end for control.From the strategic level,the autonomous motion of the omnidirectional mobile robot is taken as the starting point.By the combination of visual SLAM,ROS and Spark cloud computing platform,the autonomous control of the omnidirectional mobile robot is realized to a certain extent.