GreenWaves and Lynred Unveil Low‑Power AI Platform for Precise Building Occupancy Monitoring

Introducing a New Low‑Power AI Reference Design

GreenWaves, a leader in energy‑efficient AI processors, has partnered with infrared sensor specialist Lynred to release a reference design platform that enables building managers to optimize space utilization in connected office environments.

Why Occupancy Monitoring Matters

As companies invest in more real‑estate than necessary, real‑time occupancy data helps identify and eliminate excess space, turning unused rooms into productive assets.

Monitoring office and meeting room occupancy helps determine optimal workspace requirements (Image: Shutterstock)

The collaboration yields an evaluation board that runs on a single AA battery, accompanied by an open‑source firmware and algorithm suite. It delivers accurate person counts in spaces up to 25–30 m² while preserving occupant anonymity through exclusive infrared sensing.

AI‑Powered People Counting

“The platform’s standout feature is an embedded neural network that powers people‑counting on GreenWaves’ GAP8 processor,” said Cyrille Trouilleau, Lynred’s Smart Buildings Manager. “We trained the model on thousands of thermal images captured by Lynred’s high‑resolution 80×80 IR detector, ThermEye, across diverse office settings—from conference rooms to open‑plan layouts.”

GAP8 Processor: The Heart of the Solution

The system’s core is GreenWaves’ GAP8 application processor, engineered to execute neural‑network workloads for image, audio, radar, and vibration analysis in ultra‑low‑power IoT deployments.

The GAP8 hosts an eight‑core RISC‑V compute cluster, enhanced with GreenWaves’ custom instruction‑set extensions for extreme energy efficiency, and a dedicated control core. By isolating compute and control domains in separate voltage/frequency rails, the chip draws power only when required.

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