基于群智感知的Wi-Fi指纹室内定位精度与可靠性研究

基于Wi-Fi指纹的室内定位是移动计算领域的热点问题。现有研究着重考虑如何从系统实现方面提高定位精度，然而，Wi-Fi指纹定位技术原理可达的精度究竟如何？可靠性如何？如何从根本上提高定位性能？这些重要问题到目前为止尚未有定论，特别是近年来，群智感知技术在室内定位中被广泛应用，其带来的随机性使得回答上述问题尤为困难。为解决上述问题，本项目计划实现如下目标：建立基于群智感知的Wi-Fi指纹定位技术理论模型，刻画定位精度与可靠性相互影响的内在机理，从原理上得到定位性能极限；基于所构建的理论模型，揭示影响指纹与位置映射关系的关键因素，进而找出影响定位精度与可靠性的根本性要素；给出室内定位定标点部署特性对于定位性能的影响机理，从而找到基于群智感知的Wi-Fi指纹定位最优定标点部署策略。本项目成果对于理解基于群智感知的Wi-Fi指纹定位方法本质，进而指导定位系统设计具有借鉴意义。

Wi-Fi fingerprinting based indoor localization has been an active research field for decades. While efforts have been dedicated to improve the accuracy and efficiency of localization, some fundamental issues of the Wi-Fi fingerprinting based methodology itself is still yet studied in a theoretical perspective: how accurate the Wi-Fi fingerprinting based indoor localization can achieve? How reliable is the location estimation obtained from the Wi-Fi fingerprinting based indoor localization? How to fundamentally improve the performance of the Wi-Fi fingerprinting based indoor localization? The recently proposed solution of collecting fingerprints using the popular mobile crowdsensing makes these issues even more mysterious. . In this research, we propose a theoretical study to shed light on those fundamental issues mentioned above, with the following goals are planned to achieve: 1) Construct a general probabilistic model to describe the Wi-Fi fingerprinting based indoor localization utilizing mobile crowdsensing, in order to reveal the trade-off between localization accuracy and reliability and find the fundamental localization performance limits. 2) Find out key factors influencing the mapping between fingerprinting space and physical space in the location determination process based on the theoretical model, in order to find out key factors that influence the localization performance. 3) Reveal how the measurement points deployment for constructing the localization database could influence the localization performance, in order to find out the optimal measurement points deployment policy. The findings of the project could help understand the insight of the Wi-Fi fingerprinting based indoor localization methodology and provide guidance to the system design and implementation.