基于位置指纹的WiFi室内定位技术研究

发布时间：2018-12-13 02:20

【摘要】：随着无线通信技术的快速发展，基于位置指纹的WiFi室内定位技术以其实现简便、成本较低、定位精度高、鲁棒性好、扩展性强等优点，已经成为室内定位技术的研究热点。本文对这一技术进行了较深入和系统的研究，并针对定位准确度和异种终端差异的问题，使用迭代定位和神经网络理论提出了相应的解决方案。通过算法比较和实验分析，证明了方案的有效性和可行性，论文的主要工作如下：首先，本文通过实际WiFi环境中采集的RSSI位置指纹数据，以室内定位的角度分析了射频信号在室内的传播特性，并对可能影响定位误差的因素进行了系统的分析，为定位算法的设计与定位系统的实现提供了一定的理论依据。针对异种终端由于RSSI量化差异造成定位误差的问题，本文设计使用BP神经网络进行非线性校准作为此类问题的解决方法，使定位系统可以同时支持不同种类的终端。并使用多种终端进行数据测试，对比已有的线性校准方法，本文所设计的RSSI校准方法可以有效地提高不同终端的在已有系统下的定位精度。其次，本文以迭代定位的思想，利用定位环境中已有的非定位系统布置的AP接入点参加联合匹配定位，将定位过程分为初定位、异常感知和迭代定位三个阶段，在不增加系统复杂度的前提下，，进一步提高了原有系统的定位精度和鲁棒性。针对非布置AP位置可能发生变更的问题，设计了相对应的异常感知过程，并以情景模拟的方式验证了异常感知方法的有效性。最后，本文结合前文的理论分析和定位方法，设计并实现了一个基于位置指纹的WiFi室内定位系统原型，首先给出了室内定位系统的需求描述，然后设计了系统体系结构和通信协议，再根据体系结构中的各个部分分别进行了详细设计和实现。并在实际的WiFi无线局域网环境里对联合匹配方法进行了实验验证，实验数据表明，本文所设计的方法可以有效地改进系统的定位精度，同时降低对定位系统布置AP接入点的数量和位置因素的要求。
[Abstract]:With the rapid development of wireless communication technology, WiFi indoor location technology based on location fingerprint has become a research hotspot for its advantages such as simple implementation, low cost, high positioning accuracy, good robustness and strong expansibility. This paper makes a deep and systematic study on this technique, and puts forward the corresponding solution by using the theory of iterative localization and neural network to solve the problem of location accuracy and difference between different terminals. Through the comparison of algorithms and experimental analysis, the validity and feasibility of the scheme are proved. The main work of this paper is as follows: firstly, the RSSI position fingerprint data collected in the actual WiFi environment are used in this paper. In this paper, the propagation characteristics of RF signal in indoor are analyzed from the angle of indoor positioning, and the factors that may affect the positioning error are systematically analyzed, which provides a theoretical basis for the design of localization algorithm and the realization of positioning system. Aiming at the problem of positioning error caused by RSSI quantization difference in heterogeneous terminals, this paper designs a nonlinear calibration method using BP neural network as a solution to this problem, so that the positioning system can support different kinds of terminals at the same time. Compared with the existing linear calibration methods, the RSSI calibration method designed in this paper can effectively improve the positioning accuracy of different terminals under the existing system. Secondly, this paper uses the idea of iterative localization to take part in the joint matching localization by using the AP access points arranged by the existing non-positioning system in the localization environment. The localization process is divided into three stages: initial location, anomaly sensing and iterative positioning. Without increasing the complexity of the system, the positioning accuracy and robustness of the original system are further improved. In order to solve the problem of the possible change of the unarranged AP position, the corresponding anomaly sensing process is designed, and the effectiveness of the anomaly sensing method is verified by scenario simulation. Finally, a prototype of WiFi indoor positioning system based on location fingerprint is designed and implemented by combining the above theoretical analysis and localization method. First, the requirement description of indoor positioning system is given. Then the system architecture and communication protocol are designed and implemented in detail according to each part of the architecture. The experimental results show that the proposed method can effectively improve the positioning accuracy of the system. At the same time, the number and location of AP access points are reduced.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TN92

【参考文献】