OmniSIG for Spectrum Monitoring and Awareness in Real-Time (SMART)

DeepSig proposes the development and demonstration of a dual-use communications and sensing solution for tactical ground units, leveraging its commercial-off-the-shelf OmniSIG software suite. The solution integrates advanced Artificial Intelligence (AI) and Machine Learning (ML) algorithms with small form factor Software Defined Radios (SDRs) to enable real-time identification, classification, and localization of Signals of Interest (SOIs) in congested and contested Electromagnetic Operational Environments (EMOEs).At the core of the system is OmniSIG Engine, which isolates SOIs from background noise using instantaneous bandwidth capabilities of modern SDRs. Through robust AI/ML processing, the system performs physical-layer signal classification, significantly accelerating electromagnetic spectrum awareness. The Direction Finding (DF) capability applies angle of arrival (AoA) estimation techniques to provide bearings to enemy emitters, delivering actionable intelligence to tactical units and command elements.The neural network powering OmniSIG Engine is trained to recognize a wide range of communication protocols including push-to-talk radios, RADAR, drone control signals, Wi-Fi, and LTE. These models are developed and refined using OmniSIG Studio, DeepSig’s training environment, allowing users to annotate and train on IQ data and deploy updated models rapidly to fielded systems.The Phase I effort will focus on demonstrating the combined signal classification and basic FM communications capability on a single, tactically relevant hardware platform. This integration will highlight the feasibility and operational benefits of a unified sensing and communications solution that is hardware-agnostic, modular, and optimized for edge use. This work lays the foundation for broader integration with next-generation tactical radios and scalable deployment across military echelons.