Autonomous Reasoning Improves Performance

The US Army has awarded two contracts to Charles River Analytics (CRA) for continuing research into the applications of autonomous reasoning to improved mission performance. The service will apply CRA’s open source programming language for probabilistic modelling, FIGARO, to predict failure points in military operations systems.

Under the Army’s Energy Models of Critical Components effort (E-MC2), CRA is designing a resource monitoring, energy prediction, and decision-support tool that uses probabilistic reasoning to construct and learn models of complex, real-world systems. Using E-MC2’s models, commanders will receive notifications on possible resource consumption issues. To avoid power failures in critical situations, the Army also seeks modelling tools that provide prognostics and diagnostics for backup power equipment. Under the Probabilistic Operations Warranted for Energy Reliability Evaluation and Diagnostics (POWERED) effort, the company is using rich, modular probabilistic modelling to report on the reliability of back-up generators.

“We’re using the flexibility of FIGARO in these two efforts,” Dr. Avi Pfeffer, Chief Scientist at CRA and developer of the system, said. “FIGARO lets us quickly assemble models of interacting components that use a combination of physics-based, heuristic, and data-driven reasoning. In E-MC2, we are using FIGARO to build models of energy consumption and resource usage. In POWERED, FIGARO will provide detailed models of power generator health and status.”