Reference platform demonstrates low-power AI-based occupancy detection

AI endpoint processor startup GreenWaves has teamed up with infrared sensor manufacturer Lynred on a reference design platform for building occupancy management systems. The reference design is intended for the optimization of spaces in connected buildings such as offices.

Office management for the purposes of optimizing companies’ facilities is a growing trend, as many companies are paying for more workspace than they actually need. Monitoring the occupancy of offices and meeting rooms can help determine the optimum workspace required.


Monitoring the occupancy of offices and meeting rooms can help determine the optimum workspace required (Image: Shutterstock)

The reference design developed by Lynred and GreenWaves comes in the form of an evaluation board which runs on a single AA battery. An open source implementation of the design and the algorithm used is also available. It can accurately count the number of people in a space as big as 25-30m2, say the companies. Since it only uses infrared sensing, it also preserves occupant anonymity.

“What’s new in this reference platform is that the people counting and detection function is enabled by an embedded neural network running on the GreenWaves GAP8 processor. Sample training data is obtained using Lynred’s high-resolution 80×80 IR detector, ThermEye, which captures several thousand thermal images in office spaces from meeting rooms to open-plan areas,” said Cyrille Trouilleau, smart buildings manager at Lynred.

The design uses GreenWaves GAP8 application processor, which is designed to run neural networks for image, sound, radar and vibration processing in battery-powered IoT devices.

The processor uses a computation cluster of eight RISC-V cores with GreenWaves’ custom instruction set extensions for ultra-low power operation, plus another core for control. The compute cluster and the control core are in separate voltage and frequency domains to ensure each consumes power only when necessary.


Energy efficiency versus computing power for GreenWaves’ GAP8 processor (Image: GreenWaves)

>> This article was originally published on our sister site, EE Times Europe.

 

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.