Embedded Systems

Build an EMG game controller that tracks muscle fatigue with spectral median frequency and adapts difficulty to study rehab feedback.

Build a wearable insole system to measure gait variability, test drift detection, and turn motion data into signs of good day versus bad day patterns.

Build an ESP32 bird-call classifier with log-Mel features, model compression, and power budgeting for real field audio logging.

Build and test a smart cane tip that classifies walking surfaces with sensor data, vibration feedback, and simple machine learning.

Build an optical anemometer, analyze speckle decorrelation, and calibrate an MCU signal pipeline against a wind sensor kit.

Test encrypted backscatter links, build reader-side ML decoding, and measure how tag design affects signal recovery, noise, and security.

Benchmark TinyML inference on different MCUs, compare optimization flags, and build a regression model that predicts the best chip choice.

Build a privacy-safe road scanning system that detects lane and pothole signals, maps them, and practices edge AI, sensor fusion, and evaluation.

Build a sensor network that schedules irrigation, tracks plant stress, and measures water savings with data analysis and control testing.

Build a speech-aware amplifier and test how envelope tracking changes idle current, efficiency, and sound quality against a fixed-rail baseline.