智能轮椅室内建图及自主导航技术研究
发布时间：2018-07-21 18:03
【摘要】：随着人口老龄化压力的日益加剧以及残疾人口数目的不断增多,提高老年人和残疾人生活行动的便携性已成为急需解决的社会问题,智能轮椅作为助老助残产品中的重要分支受到国内外研究者的广泛关注。室内建图及自主导航作为智能轮椅的基础功能,其中涉及的关键技术有SLAM(Simultaneous Localization And Mapping,即同时定位与建图)、全局定位以及路径规划等。本文对智能轮椅室内建图及自主导航过程中涉及的关键技术进行研究,主要工作如下:首先,研究了SLAM的数学模型,针对基于滤波的建图算法无法在大环境中创建全局一致地图的问题,提出了一种基于图优化的建图算法。该算法采用基于相关性的扫描匹配方法实现帧间匹配,利用局部地图匹配完成回环检测,并使用位姿优化消除累计误差。通过仿真验证,该算法能够在室内大型环境中创建全局一致的地图。其次,研究了基本蒙特卡洛定位算法,针对基本蒙特卡洛定位算法应用于全局定位时存在的粒子收敛速度缓慢、样本退化及粒子贫乏的问题,提出了一种改进的蒙特卡洛定位算法。该算法将扩展卡尔曼滤波用于运动预测过程中的位姿估计,同时将自适应机制引入重采样过程,自适应地重采样及调整粒子数。通过仿真验证,该算法提高了粒子的收敛速度以及定位的可靠性。再次,研究了A*算法,针对A*算法用于全局路径规划时存在的规划效率低、规划路径转折多的问题,提出了一种双向平滑A*算法,提高了全局路径规划的效率,得到了平滑的全局路径。同时研究了DWA算法用于局部路径规划,并基于全局路径规划与局部路径规划相结合的思想,提出了一种基于改进A*算法和DWA算法的混合路径规划方法。通过仿真验证,该方法能够在已知环境中得到一条较优且平滑的全局路径,同时能够在局部变化的环境中进行避障。最后,搭建了智能轮椅实验平台,包括智能轮椅硬件平台以及基于ROS的自主导航系统。在室内复杂小环境和简单大环境中分别对智能轮椅平台进行了室内环境建图实验、定位及自主导航实验,实验结果表明本文方法具有一定的效果,同时智能轮椅能够在不同的室内环境中实现建图、定位及自主导航功能。
[Abstract]:With the increasing pressure of population ageing and the increasing number of people with disabilities, improving the portability of the lives of older persons and persons with disabilities has become a social problem that urgently needs to be addressed, As an important branch of assistive products, intelligent wheelchairs have attracted wide attention from researchers at home and abroad. Indoor mapping and autonomous navigation are the basic functions of intelligent wheelchair. The key technologies involved include slam (simultaneous localization And mapping), global positioning and path planning. The main work of this paper is as follows: firstly, the mathematical model of slam is studied. In order to solve the problem that the filter based map building algorithm can not create a global uniform map in a large environment, this paper proposes a graph building algorithm based on graph optimization. The algorithm uses correlation-based scanning matching to achieve inter-frame matching, uses local map matching to complete loop detection, and uses pose optimization to eliminate cumulative errors. Simulation results show that the algorithm can create a globally consistent map in a large indoor environment. Secondly, the basic Monte Carlo localization algorithm is studied. Aiming at the problems of the slow convergence of particles, the degradation of samples and the shortage of particles when the basic Monte Carlo localization algorithm is applied to global localization, An improved Monte Carlo localization algorithm is proposed. The extended Kalman filter is used to estimate the position and orientation of motion prediction, and the adaptive mechanism is introduced into the resampling process, which adaptively resamples and adjusts the number of particles. The simulation results show that the algorithm improves the convergence speed and the reliability of localization. Thirdly, the A * algorithm is studied. Aiming at the problems of low planning efficiency and many path turning points, a bidirectional smooth A * algorithm is proposed to improve the efficiency of global path planning. A smooth global path is obtained. At the same time, DWA algorithm is applied to local path planning. Based on the idea of combining global path planning with local path planning, a hybrid path planning method based on improved A * algorithm and DWA algorithm is proposed. The simulation results show that the proposed method can obtain an optimal and smooth global path in a known environment and can avoid obstacles in a locally changing environment. Finally, the experiment platform of intelligent wheelchair is built, including the hardware platform of intelligent wheelchair and the autonomous navigation system based on Ros. In the indoor complex environment and simple environment, the indoor environment mapping experiments, positioning and autonomous navigation experiments are carried out respectively. The experimental results show that this method has a certain effect. At the same time, intelligent wheelchair can build map, position and autonomous navigation in different indoor environment.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH789;TP242
本文编号：2136338
[Abstract]:With the increasing pressure of population ageing and the increasing number of people with disabilities, improving the portability of the lives of older persons and persons with disabilities has become a social problem that urgently needs to be addressed, As an important branch of assistive products, intelligent wheelchairs have attracted wide attention from researchers at home and abroad. Indoor mapping and autonomous navigation are the basic functions of intelligent wheelchair. The key technologies involved include slam (simultaneous localization And mapping), global positioning and path planning. The main work of this paper is as follows: firstly, the mathematical model of slam is studied. In order to solve the problem that the filter based map building algorithm can not create a global uniform map in a large environment, this paper proposes a graph building algorithm based on graph optimization. The algorithm uses correlation-based scanning matching to achieve inter-frame matching, uses local map matching to complete loop detection, and uses pose optimization to eliminate cumulative errors. Simulation results show that the algorithm can create a globally consistent map in a large indoor environment. Secondly, the basic Monte Carlo localization algorithm is studied. Aiming at the problems of the slow convergence of particles, the degradation of samples and the shortage of particles when the basic Monte Carlo localization algorithm is applied to global localization, An improved Monte Carlo localization algorithm is proposed. The extended Kalman filter is used to estimate the position and orientation of motion prediction, and the adaptive mechanism is introduced into the resampling process, which adaptively resamples and adjusts the number of particles. The simulation results show that the algorithm improves the convergence speed and the reliability of localization. Thirdly, the A * algorithm is studied. Aiming at the problems of low planning efficiency and many path turning points, a bidirectional smooth A * algorithm is proposed to improve the efficiency of global path planning. A smooth global path is obtained. At the same time, DWA algorithm is applied to local path planning. Based on the idea of combining global path planning with local path planning, a hybrid path planning method based on improved A * algorithm and DWA algorithm is proposed. The simulation results show that the proposed method can obtain an optimal and smooth global path in a known environment and can avoid obstacles in a locally changing environment. Finally, the experiment platform of intelligent wheelchair is built, including the hardware platform of intelligent wheelchair and the autonomous navigation system based on Ros. In the indoor complex environment and simple environment, the indoor environment mapping experiments, positioning and autonomous navigation experiments are carried out respectively. The experimental results show that this method has a certain effect. At the same time, intelligent wheelchair can build map, position and autonomous navigation in different indoor environment.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH789;TP242
【参考文献】
相关期刊论文 前7条
1 徐保来;管贻生;苏泽荣;谷世超;黄晶;;改进动态窗口法的阿克曼移动机器人局部路径规划器[J];机电工程技术;2016年09期
2 丁家如;杜昌平;赵耀;尹登宇;;基于改进人工势场法的无人机路径规划算法[J];计算机应用;2016年01期
3 顾青;豆风铅;马飞;;基于改进A*算法的电动车能耗最优路径规划[J];农业机械学报;2015年12期
4 潘洲;万衡;李嘉琦;;基于模糊人工势场法的移动机器人路径规划[J];制造业自动化;2015年13期
5 史恩秀;陈敏敏;李俊;黄玉美;;基于蚁群算法的移动机器人全局路径规划方法研究[J];农业机械学报;2014年06期
6 刘春明;李兆斌;黄振华;左磊;吴军;徐昕;;基于LSPI和滚动窗口的移动机器人反应式导航方法[J];中南大学学报(自然科学版);2013年03期
7 周利坤;刘宏昭;;自适应人工鱼群算法在清罐移动机器人路径规划中的应用[J];机械科学与技术;2012年07期
相关硕士学位论文 前5条
1 吴雪堂;智能轮椅检测与控制系统设计[D];沈阳工业大学;2016年
2 唐晓东;基于A*算法的无人机航迹规划技术的研究与应用[D];西南科技大学;2015年
3 王宇;面向老人的智能轮椅的研究[D];苏州大学;2015年
4 秦海春;智能轮椅自主导航行进控制技术研究[D];合肥工业大学;2014年
5 廖文龙;面向复杂环境的智能轮椅仿真及导航技术研究[D];上海交通大学;2014年
，本文编号：2136338
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2136338.html
