Design And Implementation Of Collaborative Control Experimental Platform For Multi-agent Based On Robot Operating System

In recent years,with the development of technology and artificial intelligence,the research of various kinds of robots has been deepened gradually,such as underwater robots,fire-fighting robots,rescuing robots,quadrotor.A series of robots have gradually become a hot topic of researchers.Because the single robot has limited ability to perform tasks,the research on multi-robot collaboration is deepening.Most of the research on collaborative control,searching,and detection of multiple robots is at the theoretical stage,so it is urgent to carry out the combination of relevant experiments and theories.This paper chooses a two-wheeled mobile robot to be the agent and builts a multi-intelligent experiment system based on ROS and implementes a modular software system,aiming to provide a reliable algorithm verification platform for researchers.The main work of this article is as follows:First of all,according to the requirements,the whole scheme of multi-intelligent experimental system is given.And according to the control system structure,the multi-intelligent experimental system is divided into the upper machine subsystem,the wheel mobile robot subsystem,the motion measurement subsystem and the communication subsystem.And the software and hardware design scheme of multi-intelligent system is given.Secondly,the subsystems are realized.The optitrack of existing motion measurement device of the laboratory is used to obtain the position information of the wheeled mobile robot.By using WiFi communication,wheeled mobile robots can communicate with each other.And based on event trigger,the system are integrated for entire system module,reducing the times of multi-intelligence’s communication.And based on C and ROS,this paper develops the software system,making the multi-intelligence system modular and completing the construction of hardware and software system.Finally,according to the wheel mobile robot kinematics,the nonlinear model of wheeled mobile robot is transformed into two linear second-order integrator models by dynamic feedback linearization.The formation controller is designed according to the existing formation control algorithm,and the formation control algorithm is verified by physical experiment and simulation experiment.Based on the event trigger integrated control system,the formation controller and event trigger are designed.And the designed event trigger can be effectively reducing the times of communication and it is verified by physical experiment and simulation experiment.This multi-agent collaborative control experimental platform can meet the needs of researchers for multi-agent collaborative control algorithm verification,and is committed to making the experimental platform opening,so that more researchers can build this experimental platform by themselves.And based on this platform will bring certain benefits to the development of coordinated control of multi-agent.The reason why the two-wheel differential wheeled mobile robot was selected as the controlled object of the experimental system is that the object has typical nonlinear characteristics and constitutes a complex nonlinear system,so it is the controlled object of the experimental system.