拆除机器人即时定位与地图构建算法研究

发布时间：2018-02-11 00:01

本文关键词： 拆除机器人 SLAM 多传感器信息融合定位机器人操作系ROS　出处：《安徽工业大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着城市的改造及工业的发展,拆除机器人的使用日益广泛。目前拆除机器人不具备自主移动能力,以人工遥控操作方式为主,一定程度上限制了作业效率与精度。拆除机器人的自主移动,需要借助传感器信息同时进行空间自定位与环境感知,这个过程被称为即时定位与地图构建(SLAM)。该技术涉及传感器信息处理以及多个数学模型,是拆除机器人自主移动的前提。本文首先从SLAM的一般性问题出发,阐述SLAM中定位与建图的关系;建立SLAM的概率学模型,以扩展卡尔曼滤波和粒子滤波算法为基础,对SLAM算法的具体实现进行探讨;在MATLAB中对两种SLAM算法进行仿真对比实验,以性能较好的算法作为拆除机器即时定位与地图构建的理论基础。其次,针对拆除机器人SLAM过程中的定位问题展开分具体析:探讨拆除机器人定位过程中存在的问题与难点,对拆除机器人SLAM过程中涉及的模型及两种定位方式进行了分析与建模,并在MATLAB中针对拆除机器人扫描匹配定位进行了算法验证;在此基础上,基于多传感器信息融合技术,融合里程计定位与扫描匹配定位信息,解决履带滑移带来的定位问题;以上述理论为基础,通过融合定位算法对Fast SLAM算法进行改进,作为拆除机器人即时定位与地图构建算法。最后,以上述改进的SLAM算法为基础,基于开源机器人操作系统ROS的开源包进行编码改进,在机器人仿真器Gazebo中建立仿真环境,对比里程计模型定位与融合算法定位的建图效果。在此基础上,基于拆除机器人进行硬件平台和软件平台的搭建,在实验室对改进后的算法进行实验,构建了精度较高的实验室地图,完成了对拆除机器人即时定位与地图构建算法研究。
[Abstract]:With the improvement of city and the development of industry, the demolition robot is used more and more widely. At present, the demolition robot does not have the ability to move independently, and it is mainly operated by manual remote control. To a certain extent, the efficiency and precision of the operation are limited. In order to move the robot independently, it is necessary to use the sensor information to simultaneously carry out the space self-localization and the environment perception. This process is called instant location and map building. The technology involves sensor information processing and multiple mathematical models, which is the premise of removing robot's autonomous movement. This paper starts with the general problem of SLAM. This paper expounds the relationship between location and map building in SLAM, establishes the probabilistic model of SLAM, discusses the realization of SLAM algorithm based on extended Kalman filter and particle filter algorithm, and makes a simulation and comparison experiment on two SLAM algorithms in MATLAB. The algorithm with good performance is used as the theoretical basis of the real time location and map construction of the demolition machine. Secondly, the localization problems in the SLAM process of the demolition robot are analyzed in detail: the problems and difficulties in the localization process of the demolition robot are discussed. This paper analyzes and models the models and two kinds of localization methods involved in the process of removing robot SLAM, and verifies the algorithm of scanning matching localization in MATLAB, and based on this, based on multi-sensor information fusion technology, The location problem caused by crawler slip is solved by combining the location information of mileometer and scanning, and based on the above theory, the Fast SLAM algorithm is improved by the fusion localization algorithm. Finally, based on the improved SLAM algorithm, the coding of the open source package based on the open source robot operating system ROS is improved, and the simulation environment is established in the robot simulator Gazebo. On the basis of this, the hardware platform and software platform are built based on the demolition robot, and the improved algorithm is experimented in the laboratory. The laboratory map with high precision is constructed, and the algorithms of real time localization and map construction of the dismantled robot are studied.
【学位授予单位】：安徽工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【参考文献】