Skip to:

Real-time Computing for Applications in the Field

Jeanine Cook, New Mexico State University
Muhammad Dawood, New Mexico State University
Dale Shires, Army Research Laboratory
Song-Jun Park, Army Research Laboratory
Daniel Ku, CERDEC

The core focus of Towards Real-Time Computing for Applications in the Field is to research technologies that are suitable for field deployment and applicable to problems that are important to the Army.   This project plans to focus on the development of two primary systems: a real-time imaging system for a UWB ground-penetrating radar for real-time detection of IEDs, and a real-time EEG system to monitor soldier performance in stressful field theatres.   Both of these proposed systems utilize newer computational technologies in conjunction with standard commodity CPUs.

The state-of-the-art IED detection employed by the U.S. military today comprises a downward-looking ground-penetrating radar that relies on a database of information to determine that an object is an IED.  This system costs over a million dollars and, therefore, only the most expensive ground vehicles are equipped with one.  Our goal is to create a system that costs $50,000 or less so that every unit may have at least one equipped vehicle.  This system will integrate a forward-looking ground penetrating radar (being developed primarily by ARL) and a real-time backend imaging system; the backend imaging system comprises a FPGA-CPU heterogeneous architecture.  Change detection algorithms will be used to actually detect IEDs.  An initial sweep of a transport route will be compared to the sweep during convoy transit.  Any changes detected between these two images will be flagged as a potential IED.  Because the radar is forward-looking, a vehicle can stop at a safe distance from a potential IED while traveling at a minimum of 30mph.

The portable real-time EEG system will be used primarily to monitor stress and fatigue of soldiers flying UAV missions and those in active combat situations.  The EEG system and sensors is being developed by ARL; our work focuses on a computational system that can execute AI algorithms to classify EEG signals into categories that reflect levels of stress in the soldier.  This information can subsequently be used to reposition particular personnel in battlefield situations or to determine when to rotate UAV operators.  This will promote a higher level of task efficiency and soldier safety.  The potential computational architecture comprises ARM processors and a number of Field Programmable Gate Arrays (FPGAs). The small profile/hand-held device for processing EEG in this project could potentially be applied to many other contexts of in-the-field monitoring such as data mining, imaging, and strategic planning.