DescriptionGesture recognition that enriches human-computer interaction (HCI) has gained considerable attention recently. Existing solutions such as computer-vision-based approaches recognize and track human hand/body gestures using cameras or visible light. However, they all require line-of-sight and are susceptible to interference from light sources. In this thesis, an innovative approach using ambient radio and vibration signals is implemented to achieve fine-grained hand/finger gesture recognition. By sensing the influence of hand/finger gestures on the transmitted radio signals (e.g., millimeter wave signals) and physical vibrations on a solid surface (e.g., tables, glass boards, acrylic boards), the position of the hand/finger can be precisely estimated through similarity and threshold-based techniques. Particularly, we implemented two types of solutions that work separately: (1) a mmWave-based strategy, where we leverage frequency-modulated continuous-wave (FMCW) radar to track hand movements and recognize various hand gestures, and (2) a vibration-based strategy, where we capture the tiny disturbance in the surface vibrations caused by a user’s finger touches to distinguish between different finger inputs on the surface. Extensive experiments demonstrate that our proposed approaches can accurately track and recognize user’s hand gestures with high accuracy.