基于手机传感器的室内导航定位研究

发布时间：2018-02-22 10:35

本文关键词： 小波分析多特征分析法多传感器滤波室内定位　出处：《中国地质大学(北京)》2017年硕士论文　论文类型：学位论文

【摘要】：目前导航定位市场迅速发展,在室外有已经成熟了的GNSS(Global Navigation Satellite System)提供位置服务,但是在室内无法直接接收卫星的信号,这就给位置的无盲点服务造成了很大的问题。随着现在智能手机内置传感器的快速发展,基于传感器的位置服务已经被证实在室内定位方面也有很大的潜力。该技术只需要手机内置的低精度传感器就可以实现,所以该技术具有低成本,精度高和应用广泛的前景。本文对基于手机传感器进行室内导航定位的技术进行深入研究,通过对目前室内定位研究的现状的说明,以及在传感器定位技术在室内导航定位方面可行性分析。本文做了如下几方面的工作:一、利用Eclipse进行手机传感器数据采集软件的开发,以及微电子机械系统(Micro-Electro-Mechanical System,MEMS)陀螺仪,MEMS加速度计在静止状态下的数据分析,这其中包括平稳性分析,随机误差分析。二、行进过程中手机传感器在身体不同部位的数据特点分析,这三个位置分别是脚踝、手持腰前、随臂摇摆,分析不同位置的数据特点,以及在导航定位中的应用特点,综合考虑采用了手持腰前的姿势。三、传感器数据预处理分析,分别通过IIR低通滤波、FIR低通滤波、傅里叶变换分析、小波变换对数据进行预处理,消除数据中的毛刺,使数据表现出来的波形光滑,为后面的步数计算做好准备。通过实验对比分析采用小波变化方法。四、行人步态分析以便更好的计算步数,分别使用了几种计步方法进行计步,并在以往的计步方法上面进行改进,提出多特征分析法,通过实验对比几种计步方法的准确度,采用了多特征匹配法。五、在步长方法选取方面,比较了几种成熟的步长计算方法,选取了其中的非线性步长计算方法。在航向角的选取上面,分别试验了基于磁传感器的电子罗盘方法和陀螺仪的姿态更新法,都没取得好的效果,提出了基于陀螺仪垂直轴的航向计算方法,从计算结果来看取得了很好的效果。
[Abstract]:At present, the navigation and positioning market is developing rapidly, and there is a mature GNSS(Global Navigation Satellite system outside to provide location services, but it is impossible to receive satellite signals directly indoors. This poses a big problem for location-free blind spot services. With the rapid development of smart phones with built-in sensors, Location services based on sensors have proven to have great potential in indoor positioning. The technology requires only low-precision sensors built into mobile phones, so it has a low cost. In this paper, the technology of indoor navigation and positioning based on mobile phone sensor is deeply studied, and the present situation of indoor positioning research is explained. As well as the feasibility analysis of sensor positioning technology in indoor navigation and positioning. This paper has done the following work: first, using Eclipse to develop the mobile phone sensor data acquisition software, And the data analysis of MEMS accelerometer of Micro-Electro-Mechanical system MEMS (Micro-Electro-Mechanical system MEMS) gyroscope in static state, which includes stationary analysis, random error analysis, and data characteristics analysis of mobile phone sensor in different parts of body during travel. These three positions are ankle, front of the handheld waist, swing with the arm, analyze the data characteristics of different positions, as well as the application characteristics in the navigation and positioning, and consider the posture of the front of the handheld waist synthetically. Third, the sensor data preprocessing analysis, Through IIR low-pass filter, Fir low-pass filter, Fourier transform analysis, wavelet transform to pre-process the data, eliminate the burr in the data, make the waveform of the data smooth, Through the comparative analysis of the experiment, wavelet change method is used. Fourthly, the pedestrian gait analysis is used to calculate the step number better, and several kinds of pedometer methods are used to calculate the step. The method of multi-feature analysis is put forward in this paper. By comparing the accuracy of several methods, the method of multi-feature matching is adopted. 5, in the aspect of selecting step size method, the method of multi-feature matching is used. Several mature calculation methods of step size are compared, and the nonlinear step size calculation method is selected. The electronic compass method based on magnetic sensor and the attitude updating method of gyroscope are tested on the selection of heading angle. The course calculation method based on the vertical axis of gyroscope is put forward, and the result shows that it has a good effect.
【学位授予单位】：中国地质大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP212.9

【相似文献】