【作者】 吴倩；

【导师】 刘昱；

【作者基本信息】 天津大学 ， 信息与通信工程， 2019， 硕士

【摘要】 随着科学技术的日新月异及生活方式的改变,人们对位置信息的要求越来越高。在室内环境中,由于建筑物的阻挡等因素,全球导航卫星系统(Global Navigation Satellite System,GNSS)信号会发生衰减,GNSS难以满足人们对室内定位的需求。因此,众多室内定位解决方案层出不穷,其中基于行人航位推算(Pedestrian Dead Reckoning,PDR)理论的室内定位与跟踪技术由于具备不依赖外部设备和完全自主的特点,彰显了广阔的发展前景。随着微机电系统(Micro-Electro-Mechanical System,MEMS)的发展,基于MEMS技术的惯性传感器越来越多地应用于PDR惯性定位系统中。如何在不改变制造工艺的前提下,提高惯性定位系统的精度和可靠性,成为业界关注的课题。本文以PDR技术为背景,对消费级MEMS传感器陀螺信号降噪及基于信息融合技术的人员室内定位系统的设计和实现展开了研究,主要包括以下内容:(1)MEMS陀螺信号降噪算法的研究。陀螺仪是影响MEMS传感器精度的关键部件,因此研究MEMS陀螺仪精度的提高对研究惯性定位与导航具有重要意义。本文研究了MEMS陀螺随机噪声的组成,并通过Allan分析法进行实验验证。此外,本文在研究几种非平稳、非线性信号降噪算法的基础上,提出了一种优化的降噪算法EEMD-M,并用测试信号和实测数据验证了该算法的降噪效果。(2)多传感器信息融合技术的研究。本文通过对常见的几种信息融合算法的研究,针对其尚待引入传感器观测值对权系数的影响这一因素,提出了一种基于支持度和自适应权值分配的信息融合算法,并将其用于人员室内定位系统。(3)基于PDR技术的室内定位跟踪系统的研究与实现。本课题设计、实现了一套基于经典人员惯性导航框架、结合多传感器信息融合技术的人员室内定位系统。该系统由Matlab平台、集成了多惯性测量单元(IMU)的传感器平台所构建,实现了本文所设计系统框架下的数据采集、信息融合和数据处理。通过室内环境实地测试,验证了信息融合技术的有效性,同时检验了系统的定位性能,表明本文构建的系统具有较高的实用价值。更多还原

【Abstract】 With the rapid development of science and technology and the change of lifestyle,people’s request for location information is becoming higher.In the indoor environment,the Global Navigation Satellite System(GNSS)signal will be attenuated due to the obstruction of buildings and other factors,which makes it difficult for GNSS to meet the needs of indoor positioning.Therefore,numerous indoor positioning solutions emerge one after another,among which the indoor positioning and tracking technology based on Pedestrian Dead Reckoning(PDR)demonstrates a broad development prospect for it does not rely on external facilities and is completely autonomous.With the development of Micro-Electro-Mechanical System(MEMS),inertial sensors based on MEMS technology are increasingly applied to PDR-based inertial positioning systems.How to improve the accuracy and reliability of the inertial positioning system without changing the manufacturing process has become a concerned topic in the field.Based on PDR technology,this paper studies the gyro denoising technology of consumer MEMS sensor and the design and implementation of pedestrian indoor positioning system based on information fusion.This paper mainly includes the following contents:(1)Research on MEMS gyroscope signal denoising algorithm.Gyroscope is a key component that affects the accuracy of MEMS sensor.Therefore,it is of great significance to improve the accuracy of MEMS gyroscope for the research of inertial positioning and navigation.In this paper,the composition of MEMS gyro random noise is studied and verified via Allan analysis.In addition,on the basis of studying several non-stationary and nonlinear signal denoising algorithms,this paper proposes an optimized de-noising algorithm called EEMD-M,and verifies its denoising effect with test signals and measured data.(2)Research on multi-sensor information fusion technology.Based on the research of several common information fusion algorithms,this paper proposes an information fusion algorithm based on support degree and adaptive weight allocation,for the common algorithm does not take the influence of sensor observation on weight coefficient into consideration.Besides,the fusion algorithm is applied to the human indoor positioning system.(3)Research and implementation of indoor positioning and tracking system based on PDR technology.This paper designs and implements a human indoor positioning system,which combines classical inertial navigation framework and multi-sensor information fusion technology.The system is based on the platform of Matlab and a sensor platform which is integrated with multi-inertial measurement unit(IMU).The system achieves the data acquisition,information fusion and data processing of the framework proposed in this paper.Moreover,the effectiveness of the information fusion technology and the positioning performance of the system are verified through the experiment in the indoor environment,which verifies the practical value of the system presented in this paper.更多还原