NeTS: Small: Continuous Spatial Awareness (CoSA) for Smart and Connected Objects

点击翻译按钮获取中文摘要

The Continuous Spatial Awareness (CoSA) project aims to leverage advanced signal processing techniques to allow the Wi-Fi infrastructure already deployed in the majority of workplaces, public spaces, and households to constantly sense and assess what is happening in the indoor environment. With an emphasis on availability and applicability, the project aims to support operation when just a single Wi-Fi access point is nearby and with mobiles devices that have only two to three antennas. Leveraging the proposed research techniques, the project will investigate localization of mobiles and Radio Frequency Identification (RFID) tags as well as building tomography.Embedded, smart, Internet-of-Things (IoT) objects such as radio tags, light bulbs, and wearable e-health devices are right now making their way into our daily lives. One property common to these IoT objects is that they all possess compact, convenient form factors that integrate well physically into our world. A resulting challenge is that most of these are therefore antenna-constrained: one to three antennas per device is the norm. In light of this challenge, we propose three novel methods to continuously and seamlessly track IoT objects and map our world. First, Client-Aided Localization (CAL) leverages powerful joint time- and angle-of-arrival signal processing to localize a mobile device using a single nearby access point on a single frequency band. Second, wireless building tomography maps the locations of walls indoors, thus obviating the need for a camera to visit each room of a building individually. Finally, RFID localization applies the above ideas to the domain of RFID tags with a small number of antennas. The proposed methods allow antenna-constrained devices to deduce information about their location with the cooperation of one or more antenna-rich devices nearby (such as a Wi-Fi access point). This is a modest but potentially powerful twist on previous work in angle-of-arrival and time-of-flight wireless methods that we believe will pave the way for further co-design between diverse hardware platforms at the communication end points. This research has the potential for a broad range of technology transfer opportunities to an industry which is actively engaging in the IoT. The education plan includes leveraging the research project's prototypes in a newly-created departmental independent research seminar for juniors in Computer Science.