GLIM: Generalized Detection of Low-SNR Signals Using an Iterative Feedback Model

Abstract

Accurate detection of unknown signals in low signal-to-noise ratio environments has utility in many wireless applications, such as opportunistic spectrum sharing, signal localization, and operations in long-range scenarios. Existing methods rely largely on signal processing-based techniques that perform poorly at lower energies, or machine learning techniques that rely on well-structured, offline training data with known signal labels sufficient for model training. This is impractical in environments where labeled training data is limited or difficult to obtain, such as for the detection of unknown signals that may or may not have been previously observed. To overcome these challenges, this paper introduces a novel feedback architecture for pseudo-label generation in an online-learning paradigm to detect wireless signals without a priori signal knowledge or model pre-training. The methodology improves upon digital signal processing-based techniques in low-energy detection, and performs within 3 dB of deep learning-based models trained with known signal labels, without similar limitations. The iterative architecture exhibits generalized learning as new, unknown signals are introduced to its online detection method. It is generalized for varying waveforms, sequence lengths and timing offsets, and its practical design and implementation make it ready for adoption in realistic scenarios.