【作者】 陈皎；

【导师】 盛朝强；

【作者基本信息】 重庆大学 ， 控制理论与控制工程， 2009， 硕士

【摘要】 自主机器人是一个集环境感知、动态决策与规划、行为控制与执行等多种功能于一体的综合系统,体现了信息技术与人工智能技术的发展水平。随着计算机和人工智能技术的发展,自主机器人正朝着高速、高精度、开放化、智能化、网络化发展。特别是在某些特殊的场合下,如井下作业、管道检修,若以自主机器人代替人工作业,不但能够保障人身安全、降低劳动强度,而且能够提高经济效益,给人类社会带来极大方便。自主机器人首先需要像人类一样感知周围环境,然后根据周围环境信息做出综合决策,并对自己的运动做出合理的规划,即路径规划问题。机器视觉能够为机器人提供较为丰富、较全面的环境信息,将它引入自主机器人的路径规划中,将会大大提高路径规划的效果。论文以自主机器人为研究背景,研究了以机器视觉感知导航线和障碍物,能够自主导航并避开障碍物的图像处理、目标定位、路径规划、决策控制等相关问题。论文分析研究了机器视觉图像预处理的相关算法;分析比较了多个微分算子在边缘检测中的效果,采用Roberts算子进行边缘检测,提取出导航线以及障碍物的轮廓;根据摄像机的投影模型,提出一种基于等空间间距采样的几何定位方法,利用几何关系对障碍物进行定位。为了研究路径规划算法,自主研发了履带式自主机器人。论文介绍了履带式自主机器人系统的总体设计方案;介绍了机械平台的设计流程以及加工、制作、安装、调试全过程;介绍了以DSP为核心的各个电路模块的设计方案与制作过程;搭建了完整的履带式自主机器人硬件平台;介绍了系统软件的设计思想,给出了软件模块的流程图。通过对履带式自主机器人的运动及转向数学模型的分析,依据导航线的识别以及障碍物的定位,设计了自主寻迹策略和模糊避障策略。最后,在履带式自主机器人平台上验证了路径规划算法。更多还原

【Abstract】 Autonomous robot is an integrated system with a variety of functions such as environment-aware, dynamic decision, behavior control and implementation, which embodies the development level of information technology and artificial intelligence technology. With the development of computer and artificial intelligence technology, autonomous robot is developing towards high speed, high-precision, open, intelligent and network-based. However, in some poor conditions such as mine, pipeline maintenance, the autonomous robot replacing hand manipulation is able to guarantee public safety and reduce labor intensity as well as improving economic efficiency, which brings great convenience for human society.Autonomous robot must apperceive surrounding environment information as humans first, and then make integrated decision in accordance with the surrounding environment to make a reasonable plan for their moving, namely, path planning problem. Machine vision provides richer and more comprehensive environmental information, which would improve the planning results if it is added into autonomous robot’s path plan. The thesis, on the background of autonomous robot, researches on the image processing, geometrical positioning, path planning, control strategy and so forth by employing the machine vision to perceive the navigation line and obstacles and ultimately navigating and avoiding obstacles automatically.This paper focuses on several algorithms of machine vision image pre-processing. The author adapts Roberts operator for edge detection, which extracts the navigation line and the contour of obstacles after analyzing and comparing the effect of edge detection using some differential operators. The author puts forward a method of geometrical positioning based on congruence interval sampling according to the camera’s projection model.In order to verify the algorithm under research, an autonomous tracked robot is independently developed. The thesis introduces general design of autonomous tracked robot, mechanical parts design flow and the whole process of processing, production, installation and debugging. It presents the design and producing of each circuit module based on DSP, sets up an integrated autonomous tracked robot hardware system, introduces software system design idea, and shows the flow chart of the software modules. In accordance with the analysis of the movement of autonomous tracked robot and its turning mathematical model, as well as navigation line perception and obstacles positioning, the author designs automatic tracking and fuzzy obstacle avoidance strategies.In the end, the author tests the algorithms on the autonomous tracked robot.更多还原