Smart Motion Sensor Design for Energy Saving and Surveillance

Abstract

Energy is a scared resource! Optimizing energy usage is a key attribute of our daily lives. The Pyroelectric Infrared (PIR) sensor accomplishes this goal in multiple folds. First, it optimizes energy usage by sensing user presence and applying control action to turn off appliances when human or animal absence is noticed. Second, it also provides constant surveillance against human or animal motion. Third, it is a cost-effective alternative to live camera surveillance. Finally, it is privacy-protective monitoring. The PIR sensor output crosses a predetermined threshold when human or animal presence is observed in the sensor zone. The threshold level-based hypothesis is used to apply the discrete action in controlling appliances. However, the analog output from the PIR sensor provides valuable insight into the underlying observations - movement direction, approximate distance from the sensor, fast/slow walking, and falling pattern can be estimated. Machine learning algorithms can be used to estimate parameters of interest with reasonable accuracy. In this paper, the direction of movement and distance of travel in the vicinity of the PIR sensing zone is estimated by commonly used ML algorithms. The paper depicts a unique 3D printed in-house developed wireless motion sensor node that eventually is used to create a sensor network. A group of engineering technology students designed a PIR sensor module, 3D printed housing structure and deployed it throughout the building hallway. They obtained raw data involving typical human motion from the sensor to further process it in the ML domain. This project was a part of the final year capstone design experience of a group.