Applications Of A Robotic Multi-Sensor Uxo Detection Platform

Abstract

To reduce the cost of site characterization and cleanup, it is critical to implement new technologies capable of improved detection and discrimination of Unexploded Ordnance (UXO). Digital geophysical mapping and improved physics-based discrimination have led to reduced false alarm rates. Despite these improvements, current UXO technologies still have difficulty finding small UXO that are approximately the same size as most clutter objects, and challenges remain in adequately discriminating objects during a single detection/discrimination pass. In this study, we present the results of testing a multi-sensor Electromagnetic Induction (EMI) and Ground Penetrating Radar (GPR) array system originally developed for landmine detection. Our current advanced ordnance detection system integrates sensor arrays, highly accurate global and local positioning, and on-board target detection/classification software on the front loader of a semi-autonomous robotic platform. We focus on the analysis of features from two-dimensional EMI array data collected in a controlled test pit as well as multi-sensor EMI and GPR data collected during field tests on a range. High resolution steppedfrequency synthetic aperture GPR data yield images of small UXO and supporting information for discrimination against small clutter objects. Experiments with new EMI coil configurations and timedomain processing indicate potential for enhanced discrimination.