Indoor Occupancy Detection with Heterogeneous Smart Devices: A Reinforcement Learning Approach
Abstract: The emergence of radio frequency (RF) dependent device-free indoor occupancy detection has seen slow acceptance due to its high fragility. Experimentation shows that an RF-dependent occupancy detector initially performs well in the room to be sensed. However, once the physical arrangement of objects changes in the room, the performance of the classifier degrades significantly. To address this issue, we propose a Bluetooth-dependent indoor occupancy detection system which can adapt itself in the dynamic environment. This system uses a reinforcement learning approach to predict the occupancy of an indoor environment and updates its decision policy by interacting with existing IoT devices and sensors in the room. We tested this system in three different rooms for 512 hours in total, involving three occupants. Results show that, this solution achieves 20.1% performance improvement in a dynamic environment compared to the state-of-the-art supervised learning algorithm with an average F1 score of 84.4%. This system can also predict occupancy with a maximum 83.2% F1 score in a completely unknown environment with no initial trained model.
Committee:
- Prof. Vicente Ordóñez Román, Committee Chair (CS/SEAS/UVA)
- Prof. Bradford Campbell, Advisor (CS/SEAS/UVA)
- Prof. Yuan Tian (CS/SEAS/UVA)
- Prof. Lu Feng (CS/SEAS/UVA)