基于AP RSSI特征向量相似度匹配的多尺度Wi-Fi定位方法

Multi-Scale Wi-Fi Localization Method Based on AP RSSI Feature Vector Similarity Matching

摘要: 随着Wi-Fi和智能手机的愈发普及，基于Wi-Fi来进行定位是最近越来越受关注的一个研究领域。现有针对Wi-Fi定位所开展的研究工作本质上都是一种静态的方法。针对这一问题，提出一种基于AP RSSI特征向量相似度匹配的多尺度Wi-Fi定位方法，设计了一种面向目标地点的可探知AP接收信号强度指示符(RSSI)特征向量生成机制，重点研究了相似度度量和距离度量这两个指标，并依托这两个指标实现了一种AP RSSI特征向量之间相似度匹配算法，根据相似度匹配结果提供如楼宇级、楼层级、房间级等多尺度定位。基于Android环境开展实验，实验结果表明该方法在多种尺度下均能达到96%以上的定位正确率，能够提供准确、高效及低成本的定位服务。

Abstract: Recently, Wi-Fi-based localization has gained increasing attention due to the popularity of Wi-Fi and smartphones. In terms of Wi-Fi localization, existing research is essentially static. To address this problem, a multi-scale Wi-Fi localization method based on AP RSSI feature vector similarity matching is proposed. It is designed to generate probable AP Received Signal Strength Indicator (RSSI) feature vectors based on two metrics, similarity metric and distance metric, and to implement a similarity matching algorithm between RSSI feature vectors based on these two metrics, and then to provide multi-scale localization services at the building, floor, and room levels based on similarity matching results. Experiments on the Android environment demonstrate that the localization method is capable of achieving more than 96% accuracy at different scales and is efficient, low-cost, and precise.