PetaOps/W edge-AI uProcessors: Myth or reality?

The advent of neural networks capable of learning salient features from variance in the radar data has expanded the breadth of radar applications, often as an alternative sensor or a complementary modality to camera vision. Gesture recognition for command control is probably the most commonly explored application. Nevertheless, there is a lack of suitable benchmarking datasets to assess and compare the merits of the different proposed solutions. Furthermore, most current publicly available radar datasets used in gesture recognition lack diversity or generality and are not challenging enough. We make available a unique dataset designed to meet these objectives, combining two synchronized modalities: radar and dynamic vision camera for experimenting with sensory fusion. Moreover, we propose a sparse encoding of the time domain (ADC) signals that achieve a dramatic data rate reduction (>76%) while retaining the efficacy of the downstream FFT processing (<2% accuracy loss on recognition tasks). Finally, we demonstrate early sensory fusion results based on range-Doppler maps from this radar data encoding versus the conventional approach, and dynamic vision, achieving higher accuracy than either modality alone.