UniverSense: IoT Device Pairing through Heterogeneous Sensing Signals

UniverSense: IoT Device Pairing through Heterogeneous Sensing Signals
复制标题

DOI:
10.1145/3177102.3177108
复制
发表时间:
2018-02
期刊:
Proceedings of the 19th International Workshop on Mobile Computing Systems & Applications
影响因子:
--
通讯作者:
Shijia Pan;C. Dominguez;Jun Han;Adeola Bannis;P. Tague;H. Noh;Pei Zhang
Shijia Pan;C. Dominguez;Jun Han;Adeola Bannis;P. Tague;H. Noh;Pei Zhang
中科院分区:
其他
文献类型:
--
作者:
Shijia Pan;C. Dominguez;Jun Han;Adeola Bannis;P. Tague;H. Noh;Pei Zhang

文献摘要

被引文献

相似文献

在物联网(IoT)设备之间轻松建立配对对于许多智能家居场景中的快速部署非常重要。传统的配对方法,包括密钥,QR码和RFID,通常需要特定的用户界面,表面的形状/材料,或额外的标签/阅读器。越来越多的没有接口的低资源物联网设备可能无法满足这些要求,这使得它们的配对成为一个挑战。另一方面,这些设备通常已经嵌入了用于感测任务的传感器,例如惯性传感器。这些传感器可用于与设备进行有限的用户交互,但不适合单独配对。在本文中,我们提出了UniverSense,这是一种具有惯性传感器的低资源物联网设备与配备摄像头的更强大的联网设备之间的替代配对方法。为了在它们之间建立配对,用户将低资源IoT设备移动到摄像头前面。相机和设备上的传感器都捕获低资源设备的物理运动。UniverSense将这些信号转换为一个公共状态空间,以生成用于配对的指纹。我们进行了真实世界的实验来评估UniverSense,在五名参与者进行的实验中,它达到了99.9%的F1分数。
Easily establishing pairing between Internet-of-Things (IoT) devices is important for fast deployment in many smart home scenarios. Traditional pairing methods, including passkey, QR code, and RFID, often require specific user interfaces, surface's shape/material, or additional tags/readers. The growing number of low-resource IoT devices without an interface may not meet these requirements, which makes their pairing a challenge. On the other hand, these devices often already have sensors embedded for sensing tasks, such as inertial sensors. These sensors can be used for limited user interaction with the devices, but are not suitable for pairing on their own. In this paper, we present UniverSense, an alternative pairing method between low-resource IoT devices with an inertial sensor and a more powerful networked device equipped with a camera. To establish pairing between them, the user moves the low-resource IoT device in front of the camera. Both the camera and the on-device sensors capture the physical motion of the low-resource device. UniverSense converts these signals into a common state-space to generate fingerprints for pairing. We conduct real-world experiments to evaluate UniverSense and it achieves an F1 score of 99.9% in experiments carried out by five participants.