HCC-SMALL: High-Resolution RFID Tracking With Applications In Gesture Recognition And Human-Computer Interaction

目前的手势识别方法，如基于视觉和直接感知的方法，还没有被证明是有效和实用的。同样，人机交互(HCI)方法，尽管付出了相当大的研究努力，但仍然由传统的键盘和鼠标主导。改善人机界面是一个关键的需求领域，特别是对于那些认为传统方法繁琐(充其量)或无法使用(充其量)的残疾人来说。射频识别(RFID)技术的快速发展有可能极大地推进手势识别和人机交互方面的当前挑战。RFID方法的优势包括无源标签的小型化、安全植入标签的能力以及远程跟踪多个标签的可能性；这些特征结合在一起支持基于(标记的)人体动作或手势的界面的定义、识别和开发。PI在这个项目中的目标是探索适用于最具挑战性的手势识别和人机界面应用的RFID解决方案，包括复杂的手势识别和为严重运动障碍患者提供有效的计算机接口。这些目标将通过一系列协调一致的研究、开发和评估努力来实现。首先，将利用谐波雷达方法开发RFID定位算法，并使用改进的商用组件来实现；这种方法将产生比当前功率损失技术高得多的分辨率，并将能够以具有成本效益的实施方式跟踪多个标签。接下来，将针对小工作区和大工作区进行实施。小型工作空间的实施将解决严重残疾个人基于舌头的人机界面，而大型工作空间的实施将用于手势识别和计算机控制。最后，将严格评估算法、硬件和实现的基本准确性、应用程序性能和用户效率。基本准确性评估将表征时空RFID标签跟踪的性能和限制，而人机交互和手势识别能力将通过应用程序性能和用户效率评估进行评估。更广泛的影响：这项研究从根本上解决了残疾人面临的重要接口和沟通问题，并将包括残疾人直接参与方案。项目成果将为残疾人提供更有效的接口和更强的沟通能力，从而提供更大的独立性，并最终改善社会、教育和经济机会。为了实现这些目标，该项目将为一种全新的方法奠定基础，以捕捉、表征和开发用于人机交互和手势识别应用的人体运动。其结果将在目标应用之外产生重大影响，例如在动物研究、人类行为研究，甚至安全方面(例如，监测物体的精细移动)。

Current gesture recognition efforts, such as vision and direct sensing based methods, are yet to be proven effective and practical. Similarly, human-computer interaction (HCI) methods, despite considerable research efforts, remain dominated by the traditional keyboard and mouse. Improving HCI is a critical area of need, especially for individuals with disabilities who find traditional methods cumbersome (at best) or unusable (at worst). The rapid development of radio frequency identification (RFID) technologies holds the potential to significantly advance current challenges in both gesture recognition and HCI. Advantages of RFID approaches include miniaturization of passive tags, the ability to safely implant tags, and the potential to remotely track multiple tags; these characteristics combine to support the definition, recognition, and development of interfaces based on (tagged) human movements, or gestures. The PI's goal in this project is to explore RFID solutions for the most challenging gesture recognition and HCI applications, including complex gesture recognition and effective computer interfaces for individuals with severe motor impairments. These objectives will be realized through a series of coordinated research, development, and evaluation efforts. First, RFID localization algorithms will be developed utilizing harmonic radar approaches and implemented using modified commercially available components; this approach will yield significantly higher resolution than current power loss techniques, and will be capable of tracking multiple tags in cost-effective implementations. Next, implementations for small and large workspace will be carried out. The small workspace implementation will address tongue-based HCI for individuals with severe disabilities, while the large workspace implementation will be used for hand gesture recognition and computer control. Finally, the algorithms, hardware, and implementations will be rigorously evaluated for fundamental accuracy, application performance, and user efficacy. Fundamental accuracy evaluations will characterize the performance and limits of spatial-temporal RFID tag tracking, while HCI and gesture recognition capabilities will be assessed through application performance and user efficacy evaluations. Broader Impacts: This research addresses fundamentally important interface and communications issues faced by individuals with disabilities, and will include direct programmatic involvement by individuals with disabilities. Project outcomes will afford individuals with disabilities more effective interfaces and greater communications capabilities, thereby providing greater independence and, ultimately, improved social, educational, and economic opportunities. In achieving these objectives, the project will lay the foundation for a fundamentally new approach to capturing, characterizing, and exploiting human movement for HCI and gesture recognition applications. The results will have significant impact beyond the targeted applications, for instance in animal studies, human performance studies, and even security (e.g., monitoring fine movements of objects).