Active sensing of indoor human scenarios through mobile pyroelectric infrared sensors

This work presents an active sensing technology based on PIR sensors that can identify indoor human scenarios. The goal of this research is to enhance PIR motion sensors for identifying indoor scenarios consisting of multiple static, moving human subjects. Conventional PIR sensors generate signals only when human subjects have motions within the sensor FOVs. The developed wireless PIR sensor nodes are enhanced with mobility through (1) using a servo motor or being attached to the forearm of a user. We have developed a PIR sensor signal conditioning circuit with components of programmable system on a chip (PSoC), low-noise PIR sensor signal amplifier, and servo control & position decoding. The sensory signals can be converted into a feature space: temporal correlation, spatial correlation, and intersection probability. Within the feature space, different human scenarios can be identified, and human and non-human thermal sources can be distinguished. Initial experimental results have demonstrated effectiveness of such a new sensing technology.