基于双轮平衡小车的多智能体平台设计

发布时间：2018-10-21 15:12

【摘要】：随着科技的发展,多智能体系统在无人机巡航、分布式计算、无人驾驶等领域运用得越来越广泛,多智能体控制算法的研究成为当今控制理论与控制工程专业的热点课题。本论文旨在设计一套低成本、简单可靠的多机器人物理仿真平台,供高校实验室用于多智能体控制算法的仿真与实验。首先,本文充分利用双轮自平衡小车的灵活性和轻便性,选用双轮平衡小车作为智能体机器人。论文采用拉格朗日方法建立小车的动力学模型,设计了线性二次型最优控制器,并在Matlab环境中得到了仿真结果。得到合适的控制参数后,成功将控制器移植到了基于STM32单片机的控制电路板中。其次,本文设计了一种基于颜色特征识别的多智能体定位系统。选用MV-EM120C小型Gige千兆网工业相机做图像采集,并结合计算机视觉算法库OpenCV成功实现了多智能体机器人的定位,并满足一定的可靠性和精度要求。最后,论文介绍了将多智能体算法移植到我们的小车平台上的方法,并且得到了小车的编队控制结果。实验表明,我们的多智能体小车平台能够满足实验室条件下在实时性和可靠性方面的要求。
[Abstract]:With the development of science and technology, multi-agent system has been widely used in UAV cruise, distributed computing, unmanned driving and so on. The research of multi-agent control algorithm has become a hot topic in the field of control theory and control engineering. The purpose of this paper is to design a low-cost, simple and reliable multi-robot physical simulation platform, which can be used in the simulation and experiment of multi-agent control algorithm in university laboratory. Firstly, this paper makes full use of the flexibility and portability of the two-wheel self-balancing vehicle, and chooses the two-wheel-balanced vehicle as the agent robot. In this paper, the Lagrangian method is used to establish the vehicle dynamics model, and the linear quadratic optimal controller is designed, and the simulation results are obtained in Matlab environment. After getting the appropriate control parameters, the controller is transplanted to the control circuit board based on STM32 microcontroller successfully. Secondly, a multi-agent localization system based on color feature recognition is designed. The MV-EM120C small Gige gigabit network industrial camera is used for image acquisition and the localization of multi-agent robot is successfully realized with the combination of computer vision algorithm library OpenCV and meets the requirements of reliability and precision. Finally, the method of transplanting the multi-agent algorithm to our vehicle platform is introduced, and the formation control results of the vehicle are obtained. Experiments show that our multi-agent vehicle platform can meet the requirements of real-time and reliability under laboratory conditions.
【学位授予单位】：大连理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP242

【参考文献】