AI-Assisted Twin-Sensor GPR for IED Detection

This article introduces an ultra-wideband (UWB) ground-penetrating radar (GPR) system operating at 3–6 GHz, equipped with dual sensors to improve the detection of improvised explosive devices (IEDs). The system integrates advanced beamforming, clutter suppression, and deep learning (DL) models—including W-Net, SqueezeNet, and a late-fusion Twin-sensor Modified SqueezeNet—for optimized subsurface imaging and classification. A retrieval-augmented generation (RAG) pipeline, powered by a large language model (LLM), provides real-time contextual guidance for IED defusing. The twin-sensor design reduces multipath interference and enhances feature discrimination, achieving 92.12% precision and 92.18% recall—outperforming single-sensor systems. This fusion of high-resolution GPR with AI-guided interpretation boosts situational awareness and operator safety in high-threat missions.