基于Unity3d的麦克纳姆轮移动机器人仿真平台研究

发布时间：2018-06-18 11:55

本文选题：Unity3d + 机器人　；参考：《中国矿业大学》2017年硕士论文

【摘要】：自主移动机器人目前发展速度较快,且应用领域广泛,而导航和定位是移动机器人研究的两个重要问题,所以对于机器人算法的研究一直是国内外研究的热点。在编写机器人算法时,有两种方式可以对算法进行调试,一种是将程序下载到机器人样机中,观察样机的状态;另一种是使用仿真软件建立虚拟样机,观察算法的运行效果。针对目前matlab、Webots等算法仿真软件具有的仿真效果差、使用复杂等不足,本文使用Unity3d引擎建立了一个仿真平台,仿真平台包括运动副、传感器和仿真环境,使用运动副和传感器模块可以将导入的三维模型快速建立成一个虚拟样机,并且使用随机路面和测试障碍模块可以快速建立一个虚拟仿真环境,对虚拟样机进行仿真。然后本文以麦克纳姆轮移动机器人的自主导航为例,介绍使用仿真平台进行算法仿真的方法。在仿真平台中,对A*算法进行了一定改进,将A*算法的二维空间导航改进为三维空间导航,由已知环境的导航改进为未知环境的导航,实现并优化了麦克纳姆轮移动机器人在未知三维地形环境的自主导航。最后本文建立麦克纳姆轮移动机器人的虚拟样机和三维地形环境,对改进后的算法进行仿真,同时输出仿真结果并对结果进行分析,并开展实物样机导航试验,比较了实物样机与虚拟样机的行走轨迹,验证仿真平台的可靠性。本文中搭建的仿真平台具有使用简单、仿真效果好、支持自建模型等优点,并且为机器人算法的仿真提供了一种新方法。
[Abstract]:Autonomous mobile robot is developing rapidly and widely used in many fields. Navigation and localization are two important problems in mobile robot research, so the research of robot algorithm has been a hot topic at home and abroad. There are two ways to debug the algorithm, one is to download the program to the robot prototype to observe the state of the prototype, the other is to use the simulation software to build the virtual prototype to observe the effect of the algorithm. Aiming at the shortcomings of MATLAB Webots and other algorithms simulation software, such as poor simulation effect and complex use, this paper establishes a simulation platform by using the Unity3D engine. The simulation platform includes motion pair, sensor and simulation environment. Using the motion pair and sensor module, the imported 3D model can be quickly built into a virtual prototype, and a virtual simulation environment can be built quickly by using the random road surface and test obstacle module to simulate the virtual prototype. Then, taking the autonomous navigation of McNum wheel mobile robot as an example, the method of algorithm simulation using simulation platform is introduced. In the simulation platform, the algorithm of A * is improved, the two-dimensional navigation of the algorithm is improved to the three-dimensional navigation, and the navigation of the known environment is improved to the navigation of the unknown environment. The autonomous navigation of McNum wheel mobile robot in unknown three-dimensional terrain environment is realized and optimized. Finally, the virtual prototype and 3D terrain environment of McNum wheel mobile robot are established, the improved algorithm is simulated, the simulation results are outputted and the results are analyzed, and the navigation experiment of the physical prototype is carried out. The walking tracks of the physical prototype and the virtual prototype are compared to verify the reliability of the simulation platform. The simulation platform built in this paper has the advantages of simple use, good simulation effect, self-built model and so on, and provides a new method for the simulation of robot algorithm.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】