基于WLAN位置指纹的室内定位技术研究与实现

发布时间：2018-04-29 21:24

本文选题：室内定位 + WLAN　；参考：《北京工业大学》2014年硕士论文

【摘要】：近年来，随着无线通信技术的快速发展以及移动设备的逐渐普及，各种室内环境下基于位置服务的需求变得越来越迫切。室内是人类活动最为密集且停留时间最长的场合，人们迫切地需要实现室内环境下的定位、导航以及人员物资的全方位监控等智能化泛在服务。因此，室内环境下基于位置的服务存在着大量的应用需求和广阔的应用空间。基于WLAN位置指纹的室内定位技术因其设备简单，定位精度高而成为近年来室内定位技术研究的热点。本课题对该热点进行了研究并主要完成了以下几点工作：第一，针对存在着大量干扰因素的室内环境下接收信号强度RSSI波动性较大的问题，提出了通过卡尔曼滤波算法对RSSI进行预处理，从而有效消除随机干扰，再现系统状态。通过实验确定了卡尔曼滤波算法的最佳参数值。第二，针对卡尔曼滤波算法在RSSI出现跃变时，收敛速度慢的问题，提出了一种改进的卡尔曼滤波算法，该算法能够利用前几次的RSSI观测值迅速判断出RSSI是否发生了跃变，并在有RSSI跃变发生时，修改卡尔曼滤波算法的相关参数，降低算法对发生跃变前的状态估计值的认可度，从而提高算法在发生RSSI跃变情况后的收敛速度，减小RSSI估计误差。通过实验验证了改进算法的有效性。第三，针对大范围定位时，算法的匹配运算量较大的问题，提出了一种基于区域划分和AP ID过滤的匹配算法。该算法是通过区域划分和AP ID过滤将较大定位区域内大部分参考价值较低的位置指纹过滤掉，缩小指纹匹配的范围，然后再采用最近邻法确定更精确的估计位置，从而极大地减少匹配定位过程中的运算量，提高系统的定位速度。第四，为了提高定位算法对运动目标的实时跟踪定位能力，减小定位误差，还提出了一种基于卡尔曼滤波的室内运动目标实时定位算法，该算法通过对匹配算法得到的估计位置进行卡尔滤波处理，使定位结果最大程度的接近真实的运动轨迹。并通过实验验证了该算法的有效性。第五，采用模块化的设计思想，设计并实现了一种基于WLAN位置指纹的室内定位系统验证软件，为指纹定位算法的研究和改进提供了一种平台，并加入了跃变自适应卡尔曼滤波算法和基于区域划分和AP ID过滤的匹配算法，通过该软件进行定位实验，，验证了本文所提算法的有效性，同时，也证明了该软件具有良好的定位性能。
[Abstract]:In recent years, with the rapid development of wireless communication technology and the popularity of mobile devices, the need for location-based services in various indoor environments has become more and more urgent. Indoor is the most intensive and the longest stay of human occasions, people urgently need to realize the indoor environment positioning, navigation, personnel and materials of the omnidirectional monitoring and other intelligentized ubiquitous services. Therefore, location-based services in indoor environment have a large number of application requirements and broad application space. The indoor location technology based on WLAN position fingerprint has become a hot research area in recent years because of its simple equipment and high positioning accuracy. This topic has carried on the research to this hot spot and has mainly completed the following work: Firstly, aiming at the large volatility of received signal RSSI in indoor environment with a large number of interference factors, a Kalman filter algorithm is proposed to preprocess the RSSI, which effectively eliminates the random interference and reproduces the system state. The optimal parameter value of Kalman filter algorithm is determined by experiments. Secondly, an improved Kalman filter algorithm is proposed to solve the problem of slow convergence rate of Kalman filter algorithm when the RSSI jump occurs. The algorithm can quickly determine whether the RSSI has jumped or not by using the previous RSSI observations. When the RSSI jump occurs, the relevant parameters of the Kalman filter algorithm are modified to reduce the recognition of the state estimation value before the jump, so as to improve the convergence speed of the algorithm after the RSSI jump occurs and reduce the RSSI estimation error. The effectiveness of the improved algorithm is verified by experiments. Thirdly, a matching algorithm based on region partition and AP ID filtering is proposed to solve the problem of large amount of matching operation. The algorithm filters out most of the low-reference location fingerprint in the larger location area by region partition and AP ID filtering, reduces the range of fingerprint matching, and then uses the nearest neighbor method to determine the more accurate estimated location. Thus greatly reduces the operation in the matching localization process, enhances the system localization speed. Fourthly, in order to improve the ability of real-time tracking and localization of moving targets, a real-time localization algorithm based on Kalman filter is proposed. In this algorithm, the estimated position of the matching algorithm is processed by Karl filter, so that the location results are close to the real motion trajectory as much as possible. The validity of the algorithm is verified by experiments. Fifthly, a verification software of indoor location system based on WLAN location fingerprint is designed and implemented with the modular design idea, which provides a platform for the research and improvement of fingerprint location algorithm. The adaptive Kalman filtering algorithm and the matching algorithm based on region partition and AP ID filtering are added to the algorithm. The localization experiment is carried out by the software, and the validity of the proposed algorithm is verified, at the same time, It is also proved that the software has good positioning performance.
【学位授予单位】：北京工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN925.93

【参考文献】