RF Fingerprinting with Dilated Causal Convolutions–An Inherently Explainable Architecture

Deep learning approaches have been shown to be functional approximators, with radio frequency signals being no exception. This, however, typically comes at the cost of explainability in how they arrive at answers. In this paper, we dive into our RiftNet™ algorithm that is based on dilated causal convolutions and has shown over 95% accuracy for fingerprinting 10,000 open air real-world d...