面向室内定位的WIFI探针数据预处理研究

Abstract

为验证依据多探针同时感知到的同一WIFI终端的RSSI值辨识WIFI终端是否在指定区域内的可行性，本文围绕多探针数据集的构造将将WIFI终端数据预处理流程划分为探针探测数据集解析、探针探测数据时间帧编号、面向室内人员定位的探针数据构造等三个阶段，并设计了相关预处理任务的流程并进行了实现。实验结果表明，以预处理后的数据为输入，基于BP神经网络的判别器可以以很高的准确率判别WIFI终端是否在指定的区域内，依据多探针同时感知到的同一WIFI终端的RSSI值辨识WIFI终端是否在指定区域内是可行的。 To verify the feasibility of identifying whether a WIFI terminal is within a designated indoor area based on RSSI value collected by multi WIFI detector, data preprocessing task for the construc-tion of detection data set is divided into three phase such as WIFI probe data parsing, time frame Numbering for probe data and the construction of indoor occupant location oriented data set in this paper. Flow charts for those three phases are given. In this paper, a BP neural network based discriminator for the identification that whether a WIFI terminal is within a designated indoor area is implemented with multi RSSI vector as input. And experimental results show that the precision of the discriminator for indoor area location is high. It is feasibility of identifying whether a WIFI terminal is within a designated indoor area based on multi RSSI detected.