基于Kinect深度视觉的服务机器人自定位研究

发布时间：2018-11-01 19:58

【摘要】：随着全球人口老龄化问题的出现,为了提高老人的生活质量,人们对服务机器人的需求与日俱增。目前,面向家庭的服务机器人的研究与开发已经得到了广泛的关注,家庭工作中的重复性工作或苦力工作‖越来越多的由服务机器人承担。服务机器人只有清楚自己的即时位置,以及如何从即时位置到达目标位置后,才能够有效地把特定的任务顺利完成。可见,自定位是服务机器人完成路径规划、自主导航等行为的基础,是移动机器人自主完成任务的前提。本文针对室内结构化环境,服务机器人自主定位等关键技术进行深入研究,构建了一种基于Kinect深度视觉传感器,人工信标的室内定位系统,对传统三边定位算法进行了改进。从而提高了机器人在室内结构化环境中的自主定位能力。本论文的主要内容如下:本论文首先综述了服务机器人国内外研究及应用现状,对服务机器人自定位技术进行了分析和研究,并阐述了服务机器人自定位的意义。其次,详细阐述了服务机器人硬件结构、Kinect深度视觉传感器的相关基本知识。同时构建了服务机器人坐标系模型、运动学模型。再次,针对传统边缘检测算法测量精度低和范围有限的不足,提出了基于Sobel算子的改进算法。同时,深入研究了Kinect视觉传感器测距技术和工作原理,使用Kinect深度视觉传感器进行了测距实验。实验结果验证了其测量的准确度。然后,开展了室内结构化环境下服务机器人图像配准方法研究,指出了信息熵配准方法、互信息配准方法各自的不足之处。应用边缘特征和互信息的图像配准方法,仿真实验表明,所提出的方法能够在室内结构化环境下配准人工信标。最后,对信标定位与视觉定位做了深入的研究。构建了基于Kinect深度视觉传感器和人工信标的室内定位系统,针对经典三边定位算法中容易出现的问题进行了改进与完善。将改进算法运用到构建的定位系统上,通过Matlab仿真实验,验证其自定位的可行性。
[Abstract]:With the emergence of the global aging population, in order to improve the quality of life of the elderly, the demand for service robots is increasing. At present, the research and development of home oriented service robot has been paid more and more attention. More and more repetitive work or hard work in home work are undertaken by service robot. Only when the service robot knows its immediate position and how to get to the target position from the immediate position can it effectively complete the specific task successfully. It can be seen that self-localization is the basis of the behavior of path planning and autonomous navigation of the service robot and the premise of the autonomous completion of the task of the mobile robot. In this paper, the key technologies of indoor structured environment and autonomous localization of service robot are studied in depth. An indoor positioning system based on Kinect depth vision sensor and artificial beacon is constructed, and the traditional triangulation algorithm is improved. Thus, the autonomous positioning ability of robot in indoor structured environment is improved. The main contents of this thesis are as follows: firstly, this paper summarizes the research and application of service robot at home and abroad, analyzes and studies the self-localization technology of service robot, and expounds the significance of self-localization of service robot. Secondly, the hardware structure of service robot and the basic knowledge of Kinect depth vision sensor are described in detail. At the same time, the coordinate system model and kinematics model of service robot are constructed. Thirdly, an improved algorithm based on Sobel operator is proposed to solve the problem of low accuracy and limited range of traditional edge detection algorithms. At the same time, the ranging technology and working principle of Kinect vision sensor are deeply studied, and the ranging experiment with Kinect depth vision sensor is carried out. The experimental results verify the accuracy of the measurement. Then, research on image registration method of service robot in indoor structured environment is carried out, and the shortcomings of information entropy registration method and mutual information registration method are pointed out. The method of image registration based on edge features and mutual information is used. The simulation results show that the proposed method can match artificial beacons in an indoor structured environment. Finally, the beacons location and visual positioning are deeply studied. The indoor positioning system based on Kinect depth vision sensor and artificial beacon is constructed, and the problems in the classical trilateral localization algorithm are improved and improved. The improved algorithm is applied to the constructed localization system, and the feasibility of the self-localization is verified by Matlab simulation.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.41;TP242

【参考文献】