Leveraging an Edge Based Cross-domain Service Framework to Simplify Planning

In a conflict against an advanced adversary, AOC planners are challenged by the cognitive workload to quickly plan a high-sortie rate COA that leverages all domain synergy.  Synchronized multi-domain COAs that meet commander priorities, while servicing hundreds of targets protected by advanced ISR and IADS networks is a complex task.  Metron has partnered with George Mason University to develop a Cloud Hosted Microservice (CHM) called the Edge-Based Cross-domain Service (EBCS).  First we will create AI agents for each domain and train them in their domain and multi-domain strategies in a Multi-Domain Training Environment (MDTE) tailored to the actual threats in that AOC’s geographic area.  EBCS leverages AI/RL and automation in the MDTE by providing the theater threat context and structure for machine learning at scale, utilizing Amazon’s cloud secure data enclave.    This EBCS will generate COAs, first by ingesting COCOM guidance, resources and constraints like JFC priorities, targets, blue and adversary order of battle machine-to-machine (M2M).   In Phase II we will collaborate with Kessel Run’s (KR) KRADOS Product Team to ensure our M2M interfaces meet their data formats to integrate with the KRADOS computing environment and adapt in the future as KRADOS progresses.  Additionally, planner acceptance is critical, we will integrate with KR’s software developers and AOC planners in this development to earn their trust, we have supported KR AOC projects and understand AOC requirements.   This CHM generates COAs based on the latest COCOM data flows to provide planners with the 3 highest performing COAs within 2 hours.  These COAs have a higher level of multi-domain integration than current human-derived COAs due to this MDTE training and AI automation.  The EBCS user interface enables planners to select user defined evaluation criteria, adjust COAs, and select the top COA to share with other domain planners at the 4 hour planning mark.  Our Phase I research found AOC planners could trust an AI/ML COA tool if users could select the evaluation criteria and visually see the COA in faster than real-time.  Once the air planner selects the best COA, joint planners have 2 hours to adjust their forces’ timing and request changes to their JFACC apportioned forces.  These changes are coordinated automatically via a thin-client protocol, finally, the last 2 hours of this 8 hour COA cycle are for obtaining approval from each domain’s leadership.  Sophisticated adversaries require compressed COA planning timelines, this solution will produce COAs in just 8 hours, with the domain integration required to challenge these pacing threats.  This EBCS solution can be tailored for the Oil Industry, which is also resource intensive, requires detailed planning and asset synchronization, and must adapt to changing political constraints and priorities.