Socionext and Innatera have jointly developed a human-presence detection solution that combines 60GHz FMCW radar sensing with neuromorphic Edge AI to deliver reliable, always-on detection with drastically reduced power consumption.
The solution integrates Socionext’s compact 60GHz FMCW radar, which captures detailed 3D environmental information under all lighting and weather conditions, with Innatera’s ultra-low-power Spiking Neural Processor, which interprets radar patterns directly at the sensor edge. The neuromorphic processor distinguishes human from non-human motion – including stationary individuals – achieving over 99% detection accuracy while maintaining sub-milliwatt power levels. This enables devices such as cameras and radios to remain asleep until real human presence is confirmed, extending battery life by 3–6×.
“Our 60 GHz FMCW radar delivers precise, privacy-preserving sensing in a compact form factor, making it ideal for battery-powered IoT devices,” said Matthias Neumann, Senior Marketing Manager Smart Devices & Industrial at Socionext.
“Neuromorphic computing is unlocking a new generation of always-on functions in devices, through powerful, brain-like processing of sensor data within a tiny energy footprint. Combining our Spiking Neural Processor with Socionext’s FMCW radar enables robust, real-time human-presence detection at power levels no conventional AI system can match,” said Sumeet Kumar, Innatera CEO.
