NeTS: Small: Handheld mm-Accurate Positioning for Wearables

该提案开发了一个系统，可以使用用户口袋中的手持设备以毫米（mm）精度跟踪用户身上的可穿戴设备。与过去的工作不同，它不需要外部基础设施（如摄像头或天线阵列），也不需要在身体上安装不准确、笨重和电池供电的传感器。这项提议的工作将微型射频识别（RFID）标签贴在身上的设备或用户的衣服上，这种标签很便宜（只需几美分），而且完全不需要电池。然后，它通过用户口袋里的手持RFID阅读器监控这些标签的位置。与过去使用笨重的多天线RFID读取器来定位RFID标签的解决方案不同，我们的方法只需要一个既紧凑又便携的单天线读取器。建议的工作将在商品RFID硬件和标签上全面实施和评估；该解决方案打开了一组丰富的应用程序。考虑一下像脑电图（EEG）这样的医学测试，它可以成为真正的“可穿戴”设备，头部的电极可以自动定位，而不需要医生耗时的手动测量。或者考虑一下智能纺织品，它现在可以作为基于手势的界面，跟踪用户相对于手持设备的动作。该研究提出了三个贡献：(1)提出了一种新的方法来解开从RFID标签到单天线RFID读取器穿越不同路径的无线信号-这是定位研究中的一个关键挑战。它实现了这一点，而不依赖于过去文献中使用的笨重的多天线阵列。它的核心思想是使用来自RFID标签组的信号来分离沿不同路径的无线信号组件。(2)研究了相对于已知位置的其他标签定位RFID标签的解决方案。它试图使用一种算法来做到这一点，该算法利用感兴趣的标签所经历的无线信号路径与已知位置的相邻标签之间的相似性。(3)将系统完全集成，针对两种应用：i)可穿戴EEG系统，通过附加的RFID标签跟踪头部电极。这消除了医生在进行测试之前进行繁琐的手动测量的需要，使脑电图真正便携。ii)一种内置RFID标签的智能织物，由移动设备跟踪，跟踪用户的运动，进行持续的健康监测，并创建基于手势的用户界面。

This proposal develops a system that can track wearables on a user's body with millimeter (mm) accuracy using a handheld device in the user's pocket. Unlike past work, it does so without the need for external infrastructure (e.g. cameras or antenna arrays) or inaccurate, bulky and battery-powered sensors on the body. The proposed work attaches tiny radio frequency identification (RFID) tags that are cheap (costing a few cents) and completely battery-free on to on-body devices or the user's clothing. It then monitors the locations of these tags from a handheld RFID reader in the user's pocket. Unlike past solutions that use bulky, many-antenna RFID readers to locate RFID tags, our approach needs only a single-antenna reader that is both compact and portable. The proposed work will be fully implemented and evaluated on commodity RFID hardware and tags; the solution opens up a rich set of applications. Consider medical tests such as Electroencephalogram (EEG) that can become truly 'wearable', with electrodes on the head positioned automatically without requiring time-consuming manual measurements by a physician. Or consider smart textiles that can now serve as gesture-based interfaces that track a user's movements relative to handheld devices. The proposed research presents three contributions: (1) It proposes a novel approach to disentangle wireless signals traversing different paths from RFID tags to a single-antenna RFID reader - a key challenge in positioning research. It achieves this without relying on bulky, multi-antenna arrays used in past literature. Its core idea is to use signals from teams of RFID tags to separate wireless signal components along different paths. (2) It investigates a solution to locate RFID tags relative to other tags whose positions are known. It seeks to do this using an algorithm that leverages similarity between the wireless signal paths experienced by the tag of interest and neighboring tags whose locations are known. (3) It fully integrates the system to target two applications: i) A wearable EEG system that tracks electrodes on the head using RFID tags attached to them. This eliminates the need for cumbersome manual measurements that doctors perform today before administering the test, allowing EEG to be truly portable. ii) A smart fabric with embedded RFID tags tracked by a mobile device to track the users movements both for continuous fitness monitoring and to create a gesture-based user interface.