List All Projects List Graduate Projects List Senior Design Projects List REU Projects Other Undergraduate Projects
Project # 101
Robust and Privacy-Preserving mmWave Radar Occupancy Detection for Energy-Efficient Smart Buildings
| Team Lead: | Shaun Jacob |
| Other Team Members: | |
| Faculty Advisor(s): | Dianqi Han |
| Department: | CSE |
| Project Type: | REU |
| Title: | Robust and Privacy-Preserving mmWave Radar Occupancy Detection for Energy-Efficient Smart Buildings |
| Abstract: | Buildings waste substantial energy when HVAC and lighting run in empty spaces, yet many occupancy sensors are either unreliable for stationary people or raise privacy concerns. This project develops a privacy preserving indoor occupancy detection and people counting system using 60 to 64 GHz mmWave radar. The approach targets robust presence detection at 6 to 8 m, with high accuracy for stationary occupants and reliable counting in higher density scenarios. The pipeline converts radar frames into range Doppler representations and point clouds, then suppresses static clutter using adaptive MTI filtering while preserving low frequency micro motion signatures such as breathing. DBSCAN clustering groups detections into candidate occupants, and multi target Kalman tracking enforces temporal consistency to reduce false alarms from clutter and multipath reflections. A persistence requirement over a short time window further rejects intermittent artifacts, and an auto calibration phase learns room specific noise and clutter statistics to adapt CFAR style thresholds over time. Evaluation will use multi room data with ground truth labels across 0 to 8 occupants, reporting occupancy accuracy, counting error, false alarm rate, and range performance, with comparisons to common commercial and reference baselines. |
| Poster: | View Poster #101 (PDF) |
| Judges Scoring: | Submit Evaluation (judges only) |