BrainChip has launched a Radar Reference Platform designed to provide real-time object classification at the edge. The system uses neuromorphic AI technology to identify objects that traditional radar systems can typically only locate or track by velocity.

The company stated the platform is engineered for applications where size, weight, power, and cost are restricted. Potential use cases include lightweight threat classification, fall detection in medical facilities, and obstacle detection for unmanned marine vessels.

The platform utilises Micro-Doppler signatures to distinguish between targets with similar radar returns, such as differentiating a drone from a bird. It runs on the Akida neuromorphic processor, which processes data at the point of acquisition. This on-device approach allows for real-time inference in environments without cloud connectivity and is designed for low power consumption.

The technical stack includes the BrainChip AKD1500 co-processor and an Asahi Kasei FMCW Radar Module. The software component features a pre-integrated Micro-Doppler classification model and a dashboard for visualising Range-Doppler plots. The system is targeted at sectors including defence, drone countermeasures, and contactless health monitoring.

“This platform doesn’t just see movement; it understands exactly what it is seeing, providing the mission-critical intelligence required for modern defence and autonomous systems,” said Sean Hehir, Chief Executive Officer of BrainChip.

Read more at: BrainChip Unveils Radar Reference Platform to Bridge the ‘Identification Gap’ in Edge AI – BrainChip