基于群智感知的非约束智能手机室内融合定位方法研究

项目名称： 基于群智感知的非约束智能手机室内融合定位方法研究

项目编号： No.61461037

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 无线电电子学、电信技术

项目作者： 黄宝琦

作者单位： 内蒙古大学

项目金额： 47万元

中文摘要： WiFi指纹数据库、行人航迹推算（Pedestrian Dead Reckoning，PDR）算法以及基于多源位置信息的融合定位算法是基于群智感知的WiFi指纹定位技术的重要环节，这些环节及其相互间的影响与放宽智能手机约束所带来的定位误差密切相关。本项目通过分析WiFi指纹数据库的创建机制，针对在设备异构性、位置不确定性和空间非均匀性情况下采集的WiFi指纹样本数据，采用WiFi指纹标准化、半监督机器学习和空间插值等方法探究改进的WiFi指纹数据库创建方法; 深入分析PDR算法的误差机理，揭示其在非约束条件下的传播规律，研究基于多传感器信号融合的面向非约束智能手机的鲁棒PDR算法；探索行人行为与其位置的关联规律，通过引入情境感知计算发掘行为蕴藏的位置信息，研究基于贝叶斯估计的高精度室内融合定位新算法。研究成果将为智能手机室内定位提供理论指导和实用方法，为推动和普及移动计算及其应用奠定基础。

中文关键词： WiFi指纹定位；群智感知；室内定位；智能手机；行人航迹推算

英文摘要： WiFi fingerprint databases, pedestrian dead reckoning (PDR) algorithms and fusing multiple-source location information are key to implementing crowdsourcing-based WiFi fingerprinting localization algorithms. This project will analyse the mechanism of building up WiFi fingerprint databases, and improve the process of building up WiFi fingerprint databases by standardizing WiFi fingerprints, semi-supervised machine learning and spatial interpolation given WiFi fingerprint samples from heterogeneous devices and with uncertain labels and a non-uniform spatial distribution. By exploring the error mechanism of PDR algorithms for unconstrained smart phones, a robust PDR algorithm is proposed by fusing signals from multiple Micro-Electro-Mechanical Systems (MEMS) sensors. After investigating the potential relationship between human behaviors and their location, constraints on human location can be extracted based on on context-aware computing techniques. Finally, a new and accurate indoor localization scheme based on Bayesian estimator is developed to efficiently fuse the location information from WiFi fingerprinting, PDR and context aware computing. The outcomes delivered by this project will provide theoretical guidance and practical methods for implementing indoor localization on smart phones in practice, and lay the foundation for pushing and popularizing location-based mobile computing as well as its applications.

英文关键词： WiFi Fingerprinting Localization;Crowdsourcing;Indoor Positioning;Smartphone;Pedestrian Dead Reckoning