基于粒子群优化的煤矿井下机器人FASTSLAM算法研究

发布时间：2018-04-14 12:02

本文选题：煤矿井下 + 机器人　；参考：《煤炭科学研究总院》2017年硕士论文

【摘要】：在煤矿灾害事故救援过程中,由于煤矿井下环境复杂多变,使用机器人进行侦测救援已成为现代化矿山救援技术的发展趋势,为了让机器人快速定位事故发生位置、及时侦测出灾变区域环境,必须先准确的计算出救援机器人移动轨迹及位置,在没有GPS等全球定位系统(或者无法使用GPS全球定位系统的环境)的区域,使用移动机器人同时定位与建图(Simultaneous Localization and Mapping,SLAM)技术计算机器人位置并构造井下灾害环境地图,已成为实施有效侦测、救援手段的前提。本文以中煤科工集团重庆研究院救援分院机器人实验室为主要技术支撑,瓦斯灾害监控与应急技术国家重点实验室为主要工程试验平台,采用仿真实验、现场实验相结合的方法,使用粒子群优化算法对行业现有的FASTLSAM2.0进行优化,改善算法在煤矿井下环境中的建图精度与效率,并对优化后的算法在煤矿井下环境中进行了性能实验,进行了在煤矿相似封闭环境中的仿真实验,分析影响算法性能差异的因素,主要研究内容包括以下几个方面:1.简化了机器人运动模型,使其线性化,改善了算法在位姿预测阶段的复杂度。2.针对煤矿井下的复杂环境,对激光传感器扫描出的点云地图配准方法(Iterative Closest Point,ICP),建立了优化模型,并使用罚函数法求解出其最优解,增加了点云地图配准的精度及其与环境的一致性。3.设计并编写了算法的C++代码及界面软件,对算法在实际工作环境中的效果进行了有效测试与对比,得到了基于不同环境对算法精度的影响因素分析结果。
[Abstract]:In the process of mine disaster and accident rescue, because the underground environment of coal mine is complex and changeable, the use of robot to detect and rescue has become the development trend of modern mine rescue technology.In order to detect the catastrophic environment in time, the rescue robot must first accurately calculate the trajectory and position of the robot, in areas where there is no global positioning system such as GPS (or the environment where the GPS GPS cannot be used),The use of mobile robot simultaneous location and mapping (Simultaneous Localization mapping technique) technology to calculate robot position and construct underground disaster environment map has become a prerequisite for effective detection and rescue.This paper takes the robot laboratory of Chongqing Academy of Research and Rescue Branch of China Coal Science and Technology Group as the main technical support, the State key Laboratory of Gas disaster Monitoring and Emergency Technology as the main engineering test platform, and adopts the simulation experiment.Combined with field experiments, particle swarm optimization (PSO) algorithm is used to optimize the existing FASTLSAM2.0 in the industry, to improve the accuracy and efficiency of the algorithm in the underground environment of coal mine, and the performance experiment of the optimized algorithm in the underground environment of coal mine is carried out.The simulation experiment in the similar closed environment of coal mine is carried out, and the factors influencing the performance of the algorithm are analyzed. The main research contents include the following aspects: 1.The robot motion model is simplified and linearized, and the complexity of the algorithm in the phase of pose prediction is improved.Aiming at the complex environment of underground coal mine, an optimized model is established for the point cloud map registration method scanned by laser sensor, and the penalty function method is used to solve the optimal solution.Increase the accuracy of point cloud map registration and its consistency with the environment. 3.The C code and interface software of the algorithm are designed and written, the effect of the algorithm in the actual working environment is tested and compared effectively, and the analysis results of the factors influencing the algorithm precision based on different environments are obtained.
【学位授予单位】：煤炭科学研究总院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD774

【相似文献】