Among the most common human-machine interfaces are keypads, but traditional designs are limited to discrete inputs and physically degrade over time. This project develops a contactless Analog Hall Effect Keypad using the Hall Effect along with magnetic sensing to enable continuous analog key detection. By capturing key displacement rather than simple presses, the system allows customizable actuation and dynamic mapping, providing a more versatile and expressive input interface adaptable to both industrial and consumer applications.

Birdwatching is a popular and enjoyable hobby in the United States, and enthusiasts often wonder where bird sounds are coming from and which species they belong to. This project aims to design a passive, and portable sensing module that integrates the pre-trained BirdNET model with audio processing to identify bird species and determine their locations. The system combines localization enabled by a 3D microphone array with a mobile app for user convenience.

Eye2AI is a Meta Quest 3 experience that simulates living with an English-speaking host family. Players explore a cozy home and trigger natural AI-driven conversations by gazing at everyday objects. Powered by Claude AI, the NPC host responds dynamically in spoken English, while secretly assessing the player's proficiency using the ACTFL scale. The goal is to reduce language anxiety and build real conversational fluency through immersion, not instruction.

LiDARs generate highly detailed spatial data, yet most current solutions confine this information, limiting its usefulness in dynamic, real-world scenarios. Thus, we are proud to present Hawk-Eye, an augmented reality (AR) LiDAR headset designed to seamlessly overlay spatial data directly into the user’s field of view. By combining real-time LiDAR data with accumulated point cloud maps gathered from external sources such as drones, robots, or other operators, Hawk-Eye will allow users to maintain situational awareness in unfamiliar or obstructed hazardous environments without the latency.

Returning a table tennis ball requires mass estimation, trajectory prediction, and sub-200ms mechanical response. Existing ball machines serve but cannot rally. We developed an autonomous training system featuring a paddle mounted on a three-axis servo gimbal atop a belt-driven linear gantry. Depth cameras and computer vision estimate ball position and velocity in real time, while closed-loop stepper motors and high-speed servos execute interception trajectories. The system enables solo practice with realistic, adaptive rally behavior.