A Large-scale Distribution And Deployment Of Robot Task Based On MQTT Protocol And ROS

The beginning and development of modern intelligent robots make the task of robots become more and more complex,and the large-scale distribution and deployment of robot task has become a big problem.It will cost a lot of resources,manpower and time to burn or copy each robot's task program to the robot in the traditional way.The emergence of multi-robot system promotes the development of large-scale robots,and it is urgent to distribute the same or different task programs that need to be executed by multi-robots to the robot equipment efficiently and stably.Therefore,the research of large-scale distribution and deployment platform of robot task has important theoretical significance and practical application value.The traditional distribution mechanisms mainly include Polling,SMS Push and IP Push,but for the task distribution of robot equipment,these existing distribution mechanisms cannot fully meet the needs of the system.When the network environment of the robot changes,it is also necessary to actively find a way to distribute efficiently under the limited network conditions.Therefore,this paper focuses on the large-scale distribution mechanism and mode of robot task and the task receiving and deployment mechanism of robot client.The main work of this paper can be divided into the following four parts:First of all,according to the distribution needs of scale robot task and the analysis of traditional message push technology,this paper put forward two kinds of task distribution mechanism,'direct push' and 'push-pull',based on MQTT and HTTP protocol.The communication methods of multi-robot in different network environment are investigated,and two kinds of task distribution modes,centralized and distributed,are proposed.Secondly,the robot client needs to support the receiving and running functions of the task package.According to the problem of receiving large-text robot task package,this paper provides an HTTP transmission mechanism for the receiving module,with open-sourced robot operating system ROS as the carrier,providing runtime support for the task package of robot equipment.Then,according to the implementation of the system,it adopts the B/S architecture model,and it adds the Nginx module to the Web server to achieve load balancing mechanism.It also joins the SSL/TLS communication security mechanism to the publishing server,Mosquitto,so as to improve the efficiency of the system.Finally,according to the two task distribution mechanisms proposed in this paper,the network packet analysis software Wireshark is used to test the distribution delay of the two distribution mechanisms in the same network environment.However,due to the limited experimental environment,this paper uses Docker container technology to simulate a large number of robots,and makes a more in-depth analysis of the distribution efficiency of the two distribution mechanisms.The test results show that the 'direct push' mechanism has better performance than 'push-pull' in the transmission of small text,but in the large-text task package transmission,the 'push-pull' mechanism is more efficient than the 'direct push' mechanism,and with the increase of the number and size of the robot task package,it also shows higher stability.The complete verification of operation for the entire system,using the turtlebot2 robot,selects Gmapping map function based on FastSLAM algorithm package to construct a robot task,to verify the validity of the large-scale distribution and deployment of robot task based on MQTT protocol and ROS.