AI-Enabled RF-Sensing for Radar Detection of Body-Worn IEDs

The threat posed by improvised explosive devices (IEDs) worn by suicide bombers has intensified in recent years. It is imperative to detect whether a suspected bomber is wearing an IED at a sufficiently large distance to mitigate such threats. This paper proposes an approach that utilizes radar technology and a deep-learning algorithm to identify body-worn IEDs. We employed CST Studio Suite, a high-performance 3D electromagnetic analysis software, to facilitate full-wave simulations. The fundamental characteristics of the radar cross section (RCS) of IEDs were found for various metallic, non-metallic, and human (soft tissue) models. We developed a deep learning (DL) model, which was trained and tested using data from CST simulations conducted with CST Studio Suite software. The DL model was able to achieve 96% of average detection accuracy.

Recommended citation: K. Senarathne, A. Hatharasinghe, W. Seram, D. Herath, C. Seneviratne and A. Madanayake, "AI-Enabled RF-Sensing for Radar Detection of Body-Worn IEDs," SoutheastCon 2024, Atlanta, GA, USA, 2024, pp. 644-649, doi: 10.1109/SoutheastCon52093.2024.10500269. keywords: {Radio frequency;Deep learning;Solid modeling;Radar cross-sections;Three-dimensional displays;Software algorithms;Data models;Improvised Explosive Devices;Radar Cross Section;Computer Simulation Technology},
Download Paper | Download Slides