基于foot-mounted的IMU室内行人航迹推算研究

发布时间：2018-07-12 18:02

本文选题：MEMS + IMU初始对准　；参考：《南昌大学》2013年硕士论文

【摘要】：随着现代科学技术的发展,人们对定位与导航的信息需求日益膨胀,尤其是在室内环境下,如商场、图书馆、停车场、矿井等,经常需要快速而准确地了解用户或者设备的位置信息。由于室内环境的复杂多变,电磁波传播存在衰减、多径和干扰等因素的影响,提取的信号特征在精度和稳定性方面难以得到保证。因此,基于各种无线网络(如WiFi, ZigBee等)的室内定位技术目前还不能在所有的环境中都有稳定准确的定位结果。惯性测量单元(Inertial Measure Unit, IMU)以其自主导航、不受环境干扰等优点,受到导航领域的广泛关注和重点研究。本文围绕IMU室内行人航迹推算问题进行研究,重点研究了基于微机械(Micro-Electro-Mechanical Systems, MEMS) IMU的室内行人航迹推算(Pedestrian Dead Reckoning, PDR)算法、仿真以及性能分析,主要包括以下三个方面： 1. MEMS IMU的初始对准：初始对准在导航算法中占有十分重要的地位,它的效果将直接影响导航参数的精度。由于MEMS IMU的硬件性能限制,单独采用MEMS IMU的对准精度比较差,尤其是在航向角方面。因此,本文将研究磁力计和MEMS IMU的信息融合方法,从而实现MEMS IMU的初始对准； 2.行人运动模型：IMU的运动状态是室内行人航迹推算算法中的关键问题,决定着扩展卡尔曼滤波中观测信息的准确使用。因此,本文将研究利用加速度和角速度建立行人运动模型,高效地判断IMU的运动状态。当检测到IMU处于静止状态时,采用foot-mounted方法和扩展卡尔曼滤波对导航参数误差和传感器误差进行准确的估计和及时补偿； 3.室内行人航迹推算算法：本文基于MEMS IMU硬件平台和MATLAB仿真平台,采用IMU捷联算法进行导航参数的解算,同时结合行人运动模型、foot-mounted和扩展卡尔曼滤波等方法实现误差估计和补偿,设计室内行人航迹推算算法。通过多次实验和仿真表明：位置误差占总路程的0.5%-2.0%,实现了在室内较高精度的行人航迹推算。
[Abstract]:With the development of modern science and technology, people need more and more information about positioning and navigation, especially in indoor environment, such as shopping mall, library, parking lot, mine, etc. It is often necessary to quickly and accurately understand the location of the user or device. Because of the complexity of indoor environment, the attenuation of electromagnetic wave propagation, the influence of multi-path and interference, it is difficult to ensure the accuracy and stability of the extracted signal features. Therefore, indoor localization technology based on various wireless networks (such as WiFi, ZigBee, etc.) can not have stable and accurate localization results in all environments. Inertial Measurement Unit (IMU) has been widely concerned and studied in navigation field for its advantages of autonomous navigation and non-interference with environment. This paper focuses on the research of indoor pedestrian trajectory estimation in IMU. Micro-Electro-Mechanical Systems (MEMS) IMU-based indoor pedestrian track estimation (PDR) algorithm, simulation and performance analysis are studied, including the following three aspects: 1. Initial alignment of MEMS IMU: initial alignment plays an important role in navigation algorithm, and its effect will directly affect the accuracy of navigation parameters. Because of the limitation of hardware performance of MEMS IMU, the alignment accuracy of MEMS IMU alone is poor, especially in the aspect of heading angle. Therefore, this paper will study the information fusion method of magnetometer and MEMS IMU, so as to realize the initial alignment of MEMS IMU. 2. The movement state of the pedestrian model: IMU is a key problem in the indoor footpath estimation algorithm, which determines the accurate use of the observation information in the extended Kalman filter (EKF). Therefore, in this paper, the acceleration and angular velocity are used to establish pedestrian motion model to judge the motion state of IMU efficiently. When the foot-mounted method and extended Kalman filter are used to estimate and compensate the errors of navigation parameters and sensors accurately. Indoor footpath calculation algorithm: based on MEMS IMU hardware platform and MATLAB simulation platform, the IMU strapdown algorithm is used to solve the navigation parameters. At the same time, the error estimation and compensation are realized by combining the pedestrian movement model foot-mounted and extended Kalman filter. Calculation algorithm of pedestrian track in Design Room. Through many experiments and simulations, it is shown that the position error accounts for 0.5- 2.0% of the total distance, and a high precision pedestrian track calculation is realized.
【学位授予单位】：南昌大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P227.9

【参考文献】