| dc.description.abstract | PIR sensors, known as motion detectors, are widely used for moving occupancy detection. Made of pyroelectric materials, such as LiTaO₃, generating pyroelectric current when the received infrared radiation changes, PIR sensors only respond to the motion of occupants. This results in frequent false negative detections when stationary occupancy detection is also desired, such as occupancy-based building lighting control.
To enable stationary occupancy detection, in this dissertation, we develop optical shutters to actively modulate the radiation received by the PIR sensors in the long-wave infrared (LWIR) region (8-12 µm) where human skin radiates the most. The optical shutter is made of polymer dispersed liquid crystal (PDLC) sandwiched by two germanium substrates. Each germanium substrate has an anti-reflected film on one side (the nonconductive side) to reduce the reflection.
The PDLC infrared shutter, a PIR sensor, and a driving circuit forms a synchronized low-energy electronically chopped PIR (SLEEPIR) sensor module. To better improve its performance, we devised SLEEPIR sensor nodes, and formed a SLEEPIR sensor network system with advanced machine learning algorithms.
The main contributions of this dissertation include: (i) modeling the SLEEPIR output as a function of the effective modulation, the response time of the PDLC shutter, and the time constants of the PIR sensor; (ii) quantifying the impact of the driving voltage, the mass ratio, the cell gap, and the cooling rate on the effective modulation and the response time of the PDLC shutter to obtain the optimal driving voltage and fabrication conditions that maximize SLEEPIR module’s output; and (iii) experimental validation of the SLEEPIR sensor nodes for presence detection in the lab and uncontrolled environment settings. | en |
