Waveform Design Implemented on Neuromorphic Hardware

Neuromorphic computing hardware mimics neurobiological architectures and promises eventual low power operation. Additionally, arbitrary waveform generator hardware permits the realization of complex radar waveform structures. In this paper, we combine these two technologies and investigate the potential of spiking neural networks to generate radar waveforms and their suitability in dynamic environments where adaptability is paramount. We discuss the process of development, current limitations, and critical assumptions to realizing real-time waveform adaptability with this hardware. Finally, we provide simulation and a novel application of a spiking neural network implemented on Intel's Loihi research processor to classify a waveform design dataset.