Multi-agent Platform Design Based On Two-Wheeled Balancing Robot

With the development of science and technology, multi-agent systems have great applications in UAV cruising, distributed computing and other areas, and the study on the multi-agent control algorithms is becoming a hot topic in control theory and control engineering field. This thesis aims to design a low-cost, simple and reliable physical simulation platform for the simulation of multi-agent control algorithms and verification for university laboratory.Firstly, to make full use of the flexibility and portability of the two-wheeled self-balanced vehicle, we choose it as the intelligent robot. This thesis makes use of the Lagrange equation to build the dynamic model of the robot, establishes the linear quadratic optimal controller, and obtains the simulation results in Matlab environment. After getting the right parameters, the thesis successfully implements it on the physical controller circuit board based on STM32.Secondly, we design a multi-agent positioning system based on the color feature recognition. With the help of MV-EM120C/M small GigE industrial camera and OpenCV, the localization of multi-agent robots is achieved and in the meantime certain degree of reliability and accuracy is guaranteed.Finally, the thesis introduces how the multi-agent algorithms is implemented on our platform, then successfully gets the formation control results. The experiments show that our multi-agent platform can meet laboratory requirements in real-time and reliability.