Communication and sensing techniques for smart, seamless human-environment interactions

Since its birth over 70 years ago, the computing area has gone over several generations, with significant reduction in size and cost. We are now entering a new generation of computing, called ubiquitous computing, where many small computers (smartphones, tablets, sensors, actuators, etc.) are placed on user's body or in the environments and provide many useful services to users. However, current devices and communication methods have not been able to provide fully smart and seamless interaction between users and the environments yet: either users are required to explicitly give attentions to devices and give them instructions to run, or environments need to be equipped with additional infrastructure, often expensive or inconvenient, to monitor user's presence and activities. Therefore, the next generation of computing requires devices that implicitly align with user's intention, obtain necessary information from environments as well as implicitly control them, and smart environments with minimal infrastructure setup to monitor user presence and behaviors. The goal of this research is to propose novel communication and sensing methods to enable such requirements. In particular, the proposed solutions include: (i) TextureCode, a flicker-free high-speed screen-camera communication technique to help smart glasses equipped with cameras obtain useful information from video stream on electronics displays, (ii) a body-guided communication technique that are used for authenticating users with devices and objects on every single touch interaction, (iii) EyeLight, a sensing system based on visible light to provide indoor occupancy estimation and room activity recognition services, and (iv) HandSense, an on-hand capacitive coupling-based sensing system for recognizing micro, dynamic finger gestures suitable for controlling head-mounted devices. We believe these systems provide users with more seamless interaction with surrounding environments: as the environment-user interactions implicitly align well with user's intention, data exchange, sensing, or authentication happens in the background without interruption to user's workflow.Ph.D.Includes bibliographical reference