多通道融合的家用机器人精准定位研究

发布时间：2018-09-12 13:52

【摘要】：在普适计算的大背景下,随着机器人技术的大力发展和人们生活质量的提高,家用机器人的全球市场规模正在快速扩大。在家用机器人领域,定位技术是其中最基础、最重要的一项研究内容,研究人员在家用机器人定位方面做了很多研究工作.在家用机器人的实际应用中,定位是首要的环节。为了使机器人的运动更加精确,本文主要做了以下几个方面的工作。首先,分析了家用机器人常用的室内定位技术的研究现状,并分析了多通道融合的数据融合技术的研究现状及当前融合技术的技术难点。第二,在室内WiFi定位技术方面,首先对定位的算法予以介绍,然后分析了常用的WiFi及RFID定位。在机器人的WiFi定位中,改进了一种模糊聚类融合算法的位置指纹定位法,相对于硬聚类算法,能够有效提高室内定位的定位精度。在机器人的RFID定位中,采用了将标签分类的思想,实时提高定位的效率,并对定位的误差进行了测试。第三,在运动学定位的研究中,对家用机器人的运动学进行了分析,首先分析了运动坐标系的模型理论,将运动学坐标系分为正运动模型和逆运动模型,然后对机器人的运动学定位中常用的超声波定位和里程计定位的原理进行阐述,最后在试验中完成了对里程计传感器短时间内定位的实验,并对超声波传感器系统误差的校正,并对超声波的传播距离进行了测试。在定位结果中显示,对室内定位的精度能够满足实验需求。第四,在多通道的数据融合定位中,分析了DS证据理论,针对各通道获取的数据的不确定性和证据间的相容性与互斥性,改进了一种基于二次调整加权的DS证据理论对定位的结果进行优化,在程序中设置了标志位,根据获取数据的来源判定标志位,在各通道均参与融合的情况下,将该方法与经典DS证据理论的融合结果进行比较,结果表明,基于二次调整加权的DS证据理论能够提高家用机器人的定位精度,满足室内定位的要求。
[Abstract]:Under the background of pervasive computing, with the rapid development of robot technology and the improvement of people's quality of life, the global market scale of home robot is expanding rapidly. In the field of home robot, localization technology is the most basic and the most important research content, researchers have done a lot of research work in the field of home robot location. In the practical application of home robot, positioning is the primary link. In order to make the robot motion more accurate, this paper mainly does the following work. Firstly, the research status of indoor positioning technology commonly used in home robot is analyzed, and the research status of multi-channel fusion data fusion technology and the technical difficulties of current fusion technology are analyzed. Secondly, in the aspect of indoor WiFi location technology, the location algorithm is introduced firstly, and then the commonly used WiFi and RFID localization are analyzed. In the WiFi localization of the robot, the location fingerprint location method of a fuzzy clustering fusion algorithm is improved. Compared with the hard clustering algorithm, it can effectively improve the positioning accuracy of indoor location. In the RFID localization of robot, the idea of classifying tags is adopted to improve the efficiency of localization in real time, and the error of positioning is tested. Thirdly, in the research of kinematics positioning, the kinematics of home robot is analyzed. Firstly, the model theory of kinematic coordinate system is analyzed, and the kinematics coordinate system is divided into forward motion model and inverse motion model. Then, the principle of ultrasonic positioning and mileage positioning, which is commonly used in kinematic positioning of robot, is expounded. Finally, the experiment of locating the odometer sensor in a short time is completed in the experiment, and the error correction of ultrasonic sensor system is also given. The propagation distance of ultrasonic wave was tested. The results show that the accuracy of indoor positioning can meet the experimental requirements. Fourthly, in the multi-channel data fusion positioning, the DS evidence theory is analyzed, aiming at the uncertainty of the data obtained by each channel and the compatibility and mutual exclusion among the evidence. In this paper, an improved DS evidence theory based on quadratic adjustment weights is proposed to optimize the localization results. A marker bit is set up in the program, and the symbol bit is determined according to the source of the acquired data. When all channels participate in the fusion, Compared with the fusion results of the classical DS evidence theory, the results show that the DS evidence theory based on quadratic adjustment and weighting can improve the positioning accuracy of the home robot and meet the requirements of indoor positioning.
【学位授予单位】：上海工程技术大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP242

【相似文献】