22 juillet 2020 | International, Terrestre
DARPA issues solicitation for moving-target recognition project
Andrew Eversden WASHINGTON — The U.S. Defense Department's advanced research arm issued a broad agency announcement July 15 for technology that would use algorithms to identify moving military ground vehicles. The Moving Target Recognition program from the Defense Advanced Research Projects Agency's Strategic Technology Office is a “vital part” of DARPA's “Mosaic Warfare” vision, in which each weapon system is one “tile” in a large force package that overwhelms the adversary. For the program, DARPA is interested in algorithms and collection techniques that allow synthetic aperture radar, or SAR, sensors to “detect, geolocate, and image moving ground targets,” the announcement read. If the goals of the project are met, the MTR program will then work to develop automatic target recognition algorithms for the moving target images. “Emphasis is on military vehicle targets, including slow moving vehicles whose SAR signatures are superimposed on clutter,” the announcement noted. Test for moving target recognition will include airborne data collection experiments to test and evaluate the effectiveness of algorithms. Under the contract, performers will be required to provide the airborne radar sensors and flight services, while the government team will design experiments with moving ground vehicles. DARPA anticipates handing out multiple awards. The MTR program has two phases. Phase one will focus on SAR moving target detection, geolocation and imaging, according to the announcement. It has a performance period of two years and a six-month option. Phase two, which is solicited through the July 15 notice, will center on automatic target recognition. Second phase instructions will be provided to the phase one performers before the end of the phase one base period. No award amount was provided. The U.S. Army is also working through the challenges associated with advanced target recognition capabilities, such as ensuring that algorithms receive adequate and sufficient data to mature and learn. “If you're training an algorithm to recognize cats, you can get on the internet and pull up hundreds of thousands of pictures of cats,” Gen. Mike Murray, commander of Army Futures Command, said in June. “You can't do that for a T-72 [Russian tank]. You can get a bunch of pictures, but are they at the right angles, lighting conditions, vehicle sitting camouflaged to vehicle sitting open desert?” DARPA's mosaic warfare effort includes several other projects under the Strategic Technology Office, including one that would automate aerial dogfighting. The office is also developing two complementary systems that would identify combat systems in an area available for support missions and quickly plan their route to an area. https://www.c4isrnet.com/home/2020/07/21/darpa-issues-solicitation-for-moving-target-recognition-project/