登录

  • 登录
  • 忘记密码?点击找回

注册

  • 获取手机验证码 60
  • 注册

找回密码

  • 获取手机验证码60
  • 找回
毕业论文网 > 毕业论文 > 理工学类 > 自动化 > 正文

基于ROS的室内安防机器人关键技术研究

 2022-11-28 11:11  

论文总字数:32604字

摘 要

随着科学技术的不断发展,经济的稳步增长,许多企业,园区,大型购物中心和仓库及变电所等对安全保卫工作都有大量的需求,由于巡检范围和任务的不同,人力成本持续提高,危险环境及随时变化的场景给巡逻带来了不同程度的挑战。于是,人们开始研究安防巡逻机器人,让机器人代替人完成各种的巡逻任务,解决现阶段安防领域和安全生产领域传统人工巡逻存在的弊端,以提高巡逻工作效率,降低人力成本。

室内安防机器人的技术设计多个学科的交叉,主要包括构建地图,机器人定位和导航等功能。本课题主要研究以下几个方面:(1)设计一个包括上位机和下位机的安防机器人系统。该系统用Nvidia Jetson Nano 2GB嵌入式平台作为主控制器,采用了二维激光雷达RPLIDAR A1和编码器等传感手段,完成了差分轮式移动机器人硬件平台设计。在ROS系统上综合感器感知的信息和安防任务的决策并传达给下位机,控制机器人完成任务。(2)针对搭建的硬件平台对差分轮式移动机器人进行分析并完成了基于作用和约束的运动学建模;(3)介绍基于粒子滤波方法的实时马尔可夫定位求解和基于RBPF的FAST SLAM。用MATLAB编写程序,实现了粒子滤波、扩展卡尔曼粒子滤波和扩展卡尔曼滤波,对三种算法的准确性、快速性和偏差作了对比和分析。(4)介绍了导航规划中的全局路径规划和局部路径规划,阐述了全局路径规划的A*算法和Dijkstra算法,通过仿真实现了两种算法在同一张地图中的路径规划。(5)在搭建的安防机器人上完成了室内环境的SLAM构建地图和导航规划功能的实现,并对Gmapping的粒子数和导航结果之间的关系进行了研究和调试,得出构建地图的效果不随粒子数增加而精确、构建地图的时间随粒子数增加而增加的结论。Gmapping算法在实际应用中的鲁棒性、有效性和局限性,在之后的路径规划中,完成了避障。

通过实验室室内场景的验证,在搭建的差分轮式移动机器人上完成了自主规划行走路线的安防巡逻任务并实现了对特定点位的巡逻,经过验证,该安防机器人具有巡逻效率高,运行稳定等特点。

本课题的研究成果可以为无人驾驶汽车、无人外卖车,快递机器人等实际应用的提供帮助,为今后的视觉SLAM,轨迹规划等技术学习打下了坚实的基础,在实际生活中具有深远的意义。

关键词:ROS操作系统,SLAM,路径规划 安防机器人

Abstract

With the continuous development of science and technology and the steady economic growth, many companies, parks, large shopping malls, warehouses and substations have a large demand for security work. Due to the different inspection scopes and tasks, the labor cost continues Improved, dangerous environments and changing scenes at any time have brought varying degrees of challenges to patrols. As a result, people began to study security patrol robots, allowing robots to replace humans to complete various patrol tasks, solving the drawbacks of traditional manual patrols in the security field and safety production field at this stage, so as to improve the efficiency of patrol work and reduce labor costs.

The technical design of indoor security robots crosses multiple disciplines, mainly including functions such as map building, robot positioning and navigation. This subject mainly studies the following aspects: (1) Design a security robot system including upper computer and lower computer. The system uses the Nvidia Jetson Nano 2GB embedded platform as the main controller, and uses the two-dimensional lidar RPLIDAR A1 and encoders to complete the hardware platform design of the differential wheeled mobile robot. The information sensed by the sensors and the decision-making of security tasks are integrated on the ROS system and transmitted to the lower computer to control the robot to complete the task. (2) Analyze the differential wheeled mobile robot based on the built hardware platform and complete the kinematic modeling based on effects and constraints; (3) Introduce real-time Markov positioning solution based on particle filter method and FAST SLAM based on RBPF . The program is written in MATLAB to realize particle filter, extended Kalman particle filter and extended Kalman filter. The accuracy, rapidity and deviation of the three algorithms are compared and analyzed. (4) Introduced the global path planning and local path planning in navigation planning, expounded the A* algorithm and Dijkstra algorithm of global path planning, and realized the path planning of the two algorithms in the same map through simulation. (5) Completed the realization of the SLAM map construction and navigation planning functions of the indoor environment on the built security robot, and studied and debugged the relationship between the particle number of Gmapping and the navigation results, and concluded that the effect of building the map is not random. The number of particles increases and is accurate, and the time to build a map increases with the number of particles. The robustness, effectiveness, and limitations of the Gmapping algorithm in practical applications have completed obstacle avoidance in the subsequent path planning.

The research results of this topic can provide help for practical applications such as unmanned vehicles, unmanned delivery vehicles, express robots, etc., and lay a solid foundation for future technical learning such as visual SLAM and trajectory planning, and have far-reaching in real life. significance.

Keywords: ROS robot operating system, SLAM, path planning

目 录

摘 要 I

Abstract II

第一章 绪 论 1

1.1 安防机器人的研究背景 1

1.2 安防机器人的研究现状 1

1.3 安放机器人关键的技术 2

1.4 本文的主要研究的内容 3

第二章 基于ROS的安防机器人系统设计 4

2.1 安防机器人整体结构 4

2.2 移动机器人硬件设计方案 5

2.2.1上位机系统环境设计方案 5

2.2.2下位机环境设计方案 5

2.2.3 传感器设计方案 5

2.2.4 轮子排布设计方案 6

2.3 移动机器人硬件平台设计 6

2.3.1 机械系统 6

2.3.2 传感器系统 6

2.3.3 控制执行系统 7

2.3.4 安防机器人 9

2.4 移动机器人软件平台设计 9

第三章 安防机器人系统建模 10

3.1 基于作用的运动学建模 10

3.2 基于约束的运动学建模 11

第四章 即时定位与地图构建 13

4.1 基于贝叶斯递推公式的马尔可夫定位 13

4.2 基于粒子滤波的实时马尔可夫定位求解 13

4.3 地图表示与环境地图构建 15

4.3.1 常用地图表示方法 15

4.3.2 占用栅格地图构建算法 15

4.4 基于RBPF的FAST SLAM 16

4.5 PF、EKF和EPF的MATLAB实现 16

4.6 其他SLAM 18

5.1 全局路径规划 19

5.1.1 Dijkstra算法 19

5.1.2 A*算法 19

5.2 局部路径规划 19

5.2.1 DWA算法 20

5.3 A*算法和Dijkstra算法的MATLAB实现 20

剩余内容已隐藏,请支付后下载全文,论文总字数:32604字

您需要先支付 80元 才能查看全部内容!立即支付

企业微信

Copyright © 2010-2022 毕业论文网 站点地图