Advanced Algorithms to Enhance Mobile Autonomous Robots

  • Published
  • By Shelly Rich
  • Munitions
EGLIN AIR FORCE BASE, Florida --  Air Force Research Laboratory awarded Imagination Engines, Inc., a Phase II Small Business Innovation Research contract to develop advanced algorithms for mobile autonomous robots. The Air Force needs creative terrain-sensing and multivalued behavior-fusion algorithms for these robots due to the high uncertainty and complexity of battlefield environments. Mobile autonomous robots must employ not only a variety of locomotion techniques in order to navigate different terrains, but an assortment of creative behaviors to accomplish their objectives.

The foundation of this SBIR technology is a scientific principle at least as important and fundamental as jet propulsion or nuclear energy. Specifically, it hinges on the premise that injecting critical noise levels into a system of brainstorming neural networks produces new ideas and strategies. This so-called "creativity machine paradigm" vastly outperforms genetic algorithms running on supercomputers, carrying out multidimensional optimization and invention on computational platforms common to most homes and offices. Described as the artificial intelligence domain's best bet for creating human- to transhuman-level intelligence in machines, this technology could potentially permeate all aspects of AF activities and operations, from materials discovery and autonomous weaponry to logistical planning and sensor integration.

IEI has pioneered the development of neural control systems capable of brilliant adaptation and improvisational creativity that may be used to govern truly autonomous weapons systems. The technology also enables semiautonomous military systems to perform in the event of communications loss, or to temporarily gain a clear advantage by operating outside an adversary's largely human-based OODA [observe, orient, decide, and act] loop. The technology further allows vast neural systems to automatically knit themselves into the equivalent of human brain pathways, facilitating unprecedented levels of sensor integration and unmatched machine-vision-based anomaly detection for battle damage assessment and classification (e.g., automatic target recognition).