mmWave Radar Human Pose Recognition
Date:
This was a University of Bristol software engineering group project using Texas Instruments mmWave radar.
- Built a human pose recognition pipeline with the IWRL6432BOOST sensor.
- Developed custom point-cloud preprocessing for reliable real-time motion detection.
- Implemented and tuned a 3-layer MLP classifier in PyTorch.
- Improved classification accuracy by 23.7% over baseline.
- Set up training and evaluation automation with GitHub Actions for reproducible experiments.