Deep Fusion AI RAPA: radar for robots and autonomous driving

4D radar imaging systems are rapidly emerging as a key technology for autonomous driving, offering LiDAR (Light Detection and Ranging) level performance at a fraction of the cost. Awarded in the Artificial Intelligence category, this solution was created to overcome the limitations and ineffectiveness of traditional deep learning models by ensuring high capabilities of 360-degree perception and real-time detection, as well as advanced learning of the specific characteristics of radar signals, applying optimised filtering accordingly. Trained on a proprietary dataset, RAPA accurately distinguishes between static and dynamic objects, achieving over 40 per cent more accuracy than competing solutions in some public benchmarks. Optimised for embedded edge platforms, it is a candidate as a scalable and cost-effective solution for autonomous vehicles, unmanned surface units and robotics.