In recent years, Internet-of-Thing (IoT) has experienced explosive growth and rapidly penetrated into our daily life. Sensing is one of its core function to bridge digital space and physical world. However, the market of IoT sensing applications is still in its infancy and far from maturity. In this thesis, we design and build practical systems to explore the potentials of IoT sensing. We first focus on the device-device interaction. Specifically, we design a lightweight scheme to extract physical-layer signatures of surrounding environment to secure wireless interaction between IoT devices. Second, we focus on the human-device interaction. Specifically, we exploit the sensing capability of IoTs to design two liveness detection systems for heart-sound and face authentication res...[ Read more ]

In recent years, Internet-of-Thing (IoT) has experienced explosive growth and rapidly penetrated into our daily life. Sensing is one of its core function to bridge digital space and physical world. However, the market of IoT sensing applications is still in its infancy and far from maturity. In this thesis, we design and build practical systems to explore the potentials of IoT sensing. We first focus on the device-device interaction. Specifically, we design a lightweight scheme to extract physical-layer signatures of surrounding environment to secure wireless interaction between IoT devices. Second, we focus on the human-device interaction. Specifically, we exploit the sensing capability of IoTs to design two liveness detection systems for heart-sound and face authentication respectively to avoid unauthorized access. In addition, we propose a driver hand position monitoring system to help improve the driving safety. Third, we study the spoofing issue in IoT sensing. We demonstrate attack schemes to generate adversarial ECG examples to misguide arrhythmia classification systems and adversarial wireless signals to mislead WiFi-based gesture recognition systems, which calls attention to adequate countermeasures.