Axon-Pulse present new perception paradigm for autonomous cars
Axon-Pulse revolutionizes the automotive Radar Signal Processing using Artificial Intelligent and Deep Learning techniques.
Up to now, the self-driving car hasn't reached full automation level as was expected. Key challenges revolve around sensors' performance, sensors' price point and the complexity of traffic scenarios.
DeePulse, is Axon Pulse’s next-generation radar Signal Processing engine. It uses Neural Networks, to increase automotive radar sensitivity and range, improve detection, enhance resolution & separation and improve tracking. It also enables entirely new capabilities such as: radar-AI-based targets Classification. It can classify between different road users: car, pedestrians, motorcyclist/bicyclist etc.
DeePules, is built with a pioneering approach for AI & DL radar signal processing. It takes advantage of hidden features in the radar signals to outperform classical radar algorithms and improve radar perception dramatically. Whether from Raw IQ data, through range-doppler and time-doppler representations, towards high resolution point clouds, DeePulse can use various types of radar data input, to provide comprehensive insight for any scenario.
DeePulse is a lean Neural Network SW that runs on real-time. Being software ONLY it can process data of various radars, including 4D imaging and 2D/3D cost-optimized radars.