【作者】 郭英；

【导师】 李金屏；

【作者基本信息】 济南大学 ， 计算机技术（专业学位）， 2023， 硕士

【摘要】 随着机器人领域的快速发展,使用机器人代替人类全天候地完成一些重复精度高的现象越来越普遍。机器人在结构化环境中已经有了成熟的应用,但对于具有未知、不可控因素的半结构化环境还有待探索,其关键是机器人的动态导航与避障技术。半结构化环境中移动终点、不同路况、动态障碍等因素会影响机器人的稳定运行,动态导航与避障技术能够帮助机器人动态跟踪目标、动态选取控制规则和动态避障,提高机器人的应变能力。本课题是针对半结构化环境中机器人的动态导航与避障展开技术研究,目的是提升机器人在半结构化环境中的应变能力。为了解决半结构化环境中的机器人动态导航与避障问题,本文利用俯视摄像头获取全局场景信息,制定虚拟行走路线,利用声呐及视觉传感器感知周围环境进而制定避障策略,最终通过模糊PD控制来指导机器人稳定、快速地到达指定位置。本文的具体研究内容如下:(1)提出了半结构化环境中的机器人动态导航方法。该方法利用俯视摄像头获取环境的全局信息,确定导航终点后进行动态导航。基于半结构化环境中的特点,对场景进行整体和局部建模。针对终点目标的位置发生移动的情况提出基于KCF算法的动态跟踪方法,根据面临路况的不同动态地选择模糊控制规则表,深化D控制在动态导航中的适应性,完成半结构化场景中动态导航的控制功能。(2)提出了基于视觉和声呐感知的半结构化环境中机器人实时避障算法。基于半结构化环境特点设计实时避障算法,采用摄像头和声呐环获取障碍物的方位与距离,制定避障策略,保证动态导航过程中安全性问题。本文的创新点是根据半结构化环境的特点提出多套适用于不同路况的模糊PD控制规则,实现不同路况中机器人的平稳控制。本文的研究内容是针对半结构化环境中的一些未知因素而展开的针对性研究,增强了机器人的环境适应能力,具有较高的实用价值。借助Pioneer3-DX机器人平台和MATLAB仿真平台在半结构化环境中进行充分实验,实验结果表明,所提出的动态导航与避障方法在半结构化环境中具有较好的可行性。更多还原

【Abstract】 With the rapid development of robot field,it is more and more common to use robots instead of human to complete some repetitive tasks with high precision all day long.The robot has been used in the structured environment,but the semi-structured environment with unknown and uncontrollable factors remains to be explored.The key is the dynamic navigation and obstacle avoidance technology of the robot.In a semi-structured environment,factors such as moving end points,different road conditions and dynamic obstacles will affect the stable operation of the robot.Dynamic navigation and obstacle avoidance technology can help the robot to dynamically track the target,dynamically select control rules and dynamically avoid obstacles,and improve the robot’s strain capacity.This topic is aimed at the robot dynamic navigation and obstacle avoidance technology research in semi-structured environment,the purpose is to improve the robot in the semi-structured environment of the strain ability.In order to solve the problem of robot dynamic navigation and obstacle avoidance in semi-structured environment,this thesis uses the overlooking camera to obtain global scene information,formulate virtual walking routes,uses sonar and visual sensors to perceive the surrounding environment and then formulate obstacle avoidance strategies,and finally uses fuzzy PD control to guide the robot to reach the designated position stably and quickly.The specific research content of this thesis is as follows:(1)A robot dynamic navigation method in semi-structured environment is proposed.This method uses the overhead camera to obtain the global information of the environment,and then carries on the dynamic navigation after determining the navigation end point.Based on the characteristics of semi-structured environment,model the scene as a whole and locally.A dynamic tracking method based on KCF algorithm is proposed for the moving of the destination target,and the fuzzy control rule table is dynamically selected according to different road conditions,so as to deepen the adaptability of D control in dynamic navigation and complete the control function of dynamic navigation in semi-structured scenes.(2)A real-time obstacle avoidance algorithm for robots in a semi structured environment based on visual and sonar perception has been proposed.Design a real-time obstacle avoidance algorithm based on the characteristics of a semi-structured environment,use cameras and sonar rings to obtain the orientation and distance of obstacles,formulate obstacle avoidance strategies,and ensure safety issues during the dynamic navigation process..The innovation point of this thesis is to propose several sets of fuzzy PD control rules applicable to different road conditions according to the characteristics of semi-structured environment,so as to realize the stable control of the robot in different road conditions.The research content of this thesis is targeted at some unknown factors in the semi-structured environment,which enhances the robot’s adaptability to the environment and has high practical value.With the help of the Pioneer3-DX robot platform and MATLAB simulation platform,full experiments were carried out in a semi-structured environment.The experimental results show that the proposed dynamic navigation and obstacle avoidance methods are feasible in a semi-structured environment.更多还原