基于位置指紋的WiFi室內定位技術研究
發(fā)布時間：2018-12-13 02:20
【摘要】：隨著無線通信技術的快速發(fā)展，基于位置指紋的WiFi室內定位技術以其實現簡便、成本較低、定位精度高、魯棒性好、擴展性強等優(yōu)點，已經成為室內定位技術的研究熱點。 本文對這一技術進行了較深入和系統(tǒng)的研究，并針對定位準確度和異種終端差異的問題，使用迭代定位和神經網絡理論提出了相應的解決方案。通過算法比較和實驗分析，證明了方案的有效性和可行性，論文的主要工作如下： 首先，本文通過實際WiFi環(huán)境中采集的RSSI位置指紋數據，以室內定位的角度分析了射頻信號在室內的傳播特性，并對可能影響定位誤差的因素進行了系統(tǒng)的分析，為定位算法的設計與定位系統(tǒng)的實現提供了一定的理論依據。針對異種終端由于RSSI量化差異造成定位誤差的問題，本文設計使用BP神經網絡進行非線性校準作為此類問題的解決方法，使定位系統(tǒng)可以同時支持不同種類的終端。并使用多種終端進行數據測試，對比已有的線性校準方法，本文所設計的RSSI校準方法可以有效地提高不同終端的在已有系統(tǒng)下的定位精度。 其次，本文以迭代定位的思想，利用定位環(huán)境中已有的非定位系統(tǒng)布置的AP接入點參加聯合匹配定位，將定位過程分為初定位、異常感知和迭代定位三個階段，在不增加系統(tǒng)復雜度的前提下，，進一步提高了原有系統(tǒng)的定位精度和魯棒性。針對非布置AP位置可能發(fā)生變更的問題，設計了相對應的異常感知過程，并以情景模擬的方式驗證了異常感知方法的有效性。 最后，本文結合前文的理論分析和定位方法，設計并實現了一個基于位置指紋的WiFi室內定位系統(tǒng)原型，首先給出了室內定位系統(tǒng)的需求描述，然后設計了系統(tǒng)體系結構和通信協(xié)議，再根據體系結構中的各個部分分別進行了詳細設計和實現。并在實際的WiFi無線局域網環(huán)境里對聯合匹配方法進行了實驗驗證，實驗數據表明，本文所設計的方法可以有效地改進系統(tǒng)的定位精度，同時降低對定位系統(tǒng)布置AP接入點的數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.
【學位授予單位】：電子科技大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：TN92
本文編號：2375698
[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.
【學位授予單位】：電子科技大學
【學位級別】：碩士
【學位授予年份】：2013
【分類號】：TN92
【參考文獻】
相關期刊論文 前1條
1 趙方;羅海勇;林權;馬嚴;;基于核函數法及馬爾可夫鏈的節(jié)點定位算法[J];通信學報;2010年11期
相關碩士學位論文 前2條
1 劉天舒;BP神經網絡的改進研究及應用[D];東北農業(yè)大學;2011年
2 何穎;基于WLAN室內定位系統(tǒng)的AP快速部署算法研究[D];哈爾濱工業(yè)大學;2011年
本文編號：2375698
本文鏈接：http://www.sikaile.net/kejilunwen/xinxigongchenglunwen/2375698.html
