We consider the problem of localizing a smartphone user using received signal strength (RSS) measured by a set of known network nodes in a harsh indoor environment. While the RSS of a wireless signal can be conveniently accessed, using it to estimate location is non-trivial in the presence of multipath propagation, shadowing and radio interference. Auxiliary information, such as the indoor building map and user's orientation information, potentially can help to improve localization performance. As the indoor layout is usually known as a priori, a user's moving direction or orientation in a given indoor map may contain valuable information to assist for reducing location ambiguities at estimation, typically when the radio signal channel is corrupted with noise. In this paper, we propose a double-layer hidden Markov model (DHMM) within a Bayesian learning framework for combining user orientation information and processing RSS data in the localization process to deal with RSS fluctuations induced by human body shadowing and multipath interference. Simulation and experimental results show that incorporating user orientation can potentially provide promising indoor positioning results.

中文翻译：

---

通过使用双层 HMM 结合用户方向来跟踪智能手机用户

我们考虑在恶劣的室内环境中使用由一组已知网络节点测量的接收信号强度 (RSS) 来定位智能手机用户的问题。虽然可以方便地访问无线信号的 RSS，但在存在多径传播、阴影和无线电干扰的情况下，使用它来估计位置并非易事。辅助信息，例如室内建筑地图和用户的方位信息，可能有助于提高定位性能。由于室内布局通常被称为先验，因此用户在给定室内地图中的移动方向或方位可能包含有价值的信息，以帮助减少估计时的位置模糊性，通常是在无线电信号信道被噪声破坏时。在本文中，我们在贝叶斯学习框架内提出了一个双层隐马尔可夫模型（DHMM），用于在定位过程中结合用户方向信息和处理 RSS 数据，以处理由人体阴影和多路径干扰引起的 RSS 波动。仿真和实验结果表明，结合用户方向可以提供有前景的室内定位结果。