Learning Based Autonomous Indoor Localization and Map Generation

这个CRD项目提出使用基于机器学习的指纹识别来构建一个完全自主的室内定位和测绘系统。基于位置的服务被信息和通信技术行业视为一项重要的许可技术，正在成为生活中至关重要的一部分。它在导航、跟踪、医疗保健、应急和公共安全、旅行和酒店、航空、零售和广告、石油和采矿、制造业等领域有着广泛的应用。随着物联网(IoT)传感器和设备的不断部署，随着手机和物联网市场的新理念带来新的应用场景，位置信息的价值以更快的速度成倍增长。然而，室内定位要准确地实现是具有挑战性的。传统的基于测距的技术使用无线信号的接收信号强度(RSS)、到达时间或到达角度，高度依赖于复杂且动态变化的物理传播环境。**在本项目中，我们使用位置指纹(LF)来实现室内定位。LF背后的基本思想是，信道的多径结构对于每个位置都是唯一的，并且可以被认为是位置的指纹。由于无线局域网基础设施的无处不在，我们使用WiFi的RSS作为无线电地图的指纹。我们的第一个目标是开发一个位置学习过程，可以准确地确定WiFi用户在建筑物中的位置。第二个目标是设计一种机器人启用的自动现场勘测系统，该系统可以创建WiFi无线电地图以及建筑物现场地图，而不需要人工干预和AP位置或平面图。现场勘测为LF和自动化提供了数据库，大大减少了所需的人力劳动，这是实际采用的主要障碍。该项目的预期成果将为Fortinet提供一个工作原型，可以转移到他们的产品线上。所开发的技术也可被加拿大高科技行业采用，以增加其产品供应并造福最终用户。****************

This CRD project proposes to build a fully autonomous indoor localization and mapping system using machine learning based fingerprinting. Location based services are considered by the ICT industry as a significant permissive technology and becoming a vital part of life. It has wide applications in navigation, tracking, healthcare, emergency and public safety, travel and hospitality, aviation, retail and advertising, oil and mining, manufacturing, etc. With the growing deployment of Internet-of-Things (IoT) sensors and devices, the value of location information multiplies even faster as new ideas in the mobile phone and IoT markets bring new application scenarios. Indoor localization, however, is challenging to realize with accuracy. Conventional ranging based techniques, using the received signal strength (RSS), time of arrival or angle of arrival of wireless signals, are highly dependent on the complex and dynamically changing physical propagation environment. ** In this project, we use location fingerprinting (LF) to realize indoor positioning. The basic idea behind LF is that the multipath structure of the channel is unique to every location and can be considered as a fingerprint of location. Due to the ubiquity of wireless local area network infrastructure, we use the RSS of WiFi as the fingerprint of a radio map. Our first objective is to develop a location learning process that can accurately determine the location of a WiFi user in a building. The second objective is to design a robot enabled automatic site survey system that creates a WiFi radio map as well as the building site map without human intervention and the need for AP location or floor plan. The site survey provides the database for LF and automation greatly reduces the human effort needed which is a major hurdle to practical adoption. The expected outcome of the project will provide a working prototype to Fortinet, which can be transferred to their product line. The technologies developed can also be adopted by Canadian high-tech industries to enhance their product offering and benefit end users. ****************