Research On Real-time Communication Of Cloud Robot Based On Pose Estimation

With the rapid development of robot and robot vision,it is very important to improve the limited computing capacity and storage capacity of the robots.Because of the advantages of cloud computing in computing capacity and storage capacity,as well as the advantages of modularization and robustness of system application,cloud robot is called an emerging hotspot research field in robotics research field.In this thesis,a robot cloud communication system is studied and designed,which is based on monocular visual odometer,and the research of cloud robot communication and ROS robot operating system.In the indoor scene of the robots,we use the monocular visual odometry algorithm to obtain the rotation motion information and translation motion information of the robot through the color image sequence.Visual odometry is an important algorithm to obtain the pose information of robot motion.There is a positive correlation between the accuracy of the algorithm and the time complexity of the algorithm,and it has a negative correlation with the real-time performance of the algorithm.In order to improve the real-time performance of the algorithm,we use ORB feature matching algorithm as the feature processing algorithm.This algorithm can greatly improve the real-time performance of the system while guaranteeing the accuracy requirement.After obtaining the pose information of the robot,we set the threshold value of the robot pose motion,respectively,which include the robot rotation and translation information.The thresholds are obtained by weighting the rotation and translation to achieve the validity and accuracy of the key frame extraction in the image sequence.In the aspect of cloud robot communication,the communication protocol and the encoding and decoding of the sensor information and the commonness of the data format are considered in this research,and the communication protocol based on RosBridge is proposed.This kind of communication protocol is a kind of comprehensive utilization of Socket communication protocol and Websocket communication protocol.This kind of communication protocol using the perfect compatibility between RosBridge and ROS robot operating system under the premise of ensuring the smooth transmission of sensor information,which guarantees the general performance of sensor information.At last,a real-time communication system of cloud robot is built.The system includes a robot hardware system and a cloud computing system based on AWS Amazon cloud services.The robotic hardware system uses Kinect to obtain the RGB-D image sequence information in the robot motion scene through motion control of iRobot Create robot.And then the system use the notebook computer to achieve the key frame image extraction.Finally,the image is transmitted to EC2 elastic cloud of Amazon cloud service by methods of RosBridge-based communication.At the end of this thesis,the research work is summarized,and the future research work is put forward.