Operationalized Machine Intelligence Planning Assistant (OMIPA)
Navy SBIR FY2018.1


Sol No.: Navy SBIR FY2018.1
Topic No.: N181-018
Topic Title: Operationalized Machine Intelligence Planning Assistant (OMIPA)
Proposal No.: N181-018-0053
Firm: Perceptronics Solutions, Inc.
3527 Beverly Glen Blvd.
Sherman Oaks, California 91423
Contact: Timur Chabuk
Phone: (571) 235-5720
Web Site: http://www.percsolutions.net
Abstract: This proposal is to develop an innovative Operationalized Machine Intelligence Planning Assistant (OMIPA) system for next generation mission planning. The goal of the OMIPA system is to revolutionize the air mission and strike planning process through the application of advanced machine learning and artificial intelligence. Mission and strike planning is a difficult and time-consuming process. Given a variety of mission objectives, the space of possible mission plans to achieve those objectives is too large for human planners to exhaustively explore. OMIPA will revolutionize the mission and strike planning process through the integration of machine learning and artificial intelligence techniques. Technically, OMIPA will feature adaptations of cutting edge AI techniques, including case based analysis, semi-supervised learning, adversarial RL training of neural networks, Monte Carlo tree search, and evolutionary computation. These techniques will help human planners create better, highly robust plans more quickly and reliably for a range of missions; they have been chosen to avoid the technical challenges described above. Powerful ML and AI algorithms will drive OMIPAƒ?Ts ability to generate actionable insights in the mission planning process, but it is our staged approach to the development and implementation of OMIPA that will be the key to system success,
Benefits: Strike operations involve mutually supportive players who must collect intelligence to support planning, suppress enemy air defense, enable and route the strike elements to maximize their probability of successful ingress, weapon delivery, egress, and results assessment. This requires that the right assets are available and in position, and that they execute their responsibilities in synchronization. There is an exciting opportunity for artificial intelligence methods such as machine learning to provide automated support for the mission and strike planning process. Recent studies have demonstrated tremendous capabilities for machine learning techniques to outperform human beings in varied tasks. These same techniques can be applied to mission and strike planning, and indeed, it is widely believed that adversaries of the United States are actively engaged in developing such methods. OMIPA will provide a rational, incremental approach to operationalizing artificial intelligence support for mission and strike planning that explicitly recognizes and addresses these challenges. Recent advances in machine learning and artificial intelligence are poised to revolutionize the mission and strike planning process across many problems. But only a thoughtful, practical, and incremental approach that is respectful of the existing planning process can hope to successfully operationalize these capabilities. We believe OMIPA will do exactly, will be a pioneering development in the radical shift in mission planning technology, and will significantly affect future development as well.

Return