LG-based Explainable AI to Enable Realistic Theater-level Naval Wargaming

Without representing targets and threats from Air, Land, Sea Surface as well as sensors from all the domains (including Space), a system attempting to capture the TUSW operations would not be able to automatically generate Courses of Actions (COAs) that are necessary to conduct and automate wargaming of that includes the Undersea domain.As the variety and size, as well as frequency of emergence of the multidomain forces and threats increases, unaided decision-making as well as manual manipulation of the wargaming entities, makes it virtually impossible to effectively respond to wargaming events. Further, this complexity can quickly overwhelm the wargame human controller’s ability to coordinate between subsurface and other forces and to optimally allocate assets. This will sharply reduce the benefits of the wargaming to the operational command. Finally, the AI used to automate COA generation and decision-making, needs to be explainable in the sense that this AI must possess an intrinsic ability to explain the reasons for both the generated COAs and recommended actions in terms that are both tactically accurate and intuitively clear to the military personnel. To address these problems, Linguistic Geometry (LG) based explainable AI1 provides predictive, real-time, dynamic kill-web based decision-making, mission planning, battle management, and post-mission analysis capabilities based on automatic tactically accurate COA generation. This AI-based capability is fast, flexible, scalable, and explainable – and it applies to multi-domain, multi-sensor, and multi-weapon battlespaces.Over the last 25 years, STILMAN’s software was successfully tested and improved in more than 50 defense-related projects. They include three projects with DARPA, multiple SBIR Phases I, II and III projects with the Army, Navy, MDA, and Air Force, as well as projects with Dstl (UK) and various military contractors. Currently, STILMAN is the key AI participant in several large projects on the cutting-edge of technology, including “Gen. Robert B. Neller Center for Wargaming and Analysis” (PIONEER WGC) with USMC (2020-) and MINERVA (for Austere Challenge AC24 and beyond) with USAREUR (2023-2024) (both under the BAE Systems (US) leadership), and several SBIR Phases II and III projects with MDA and NAVSEA – all leading to the transition of LG-RAID to the US Armed Forces.In the Phase I of this project and, if successfully awarded, Phase II, STILMAN intends to develop the capability for LG-RAID to model and simulate such tactical tasks for submarine operations as sensing, tracking hostile subs/fleets, and destroying enemy vessels corresponding to actual/sanitized capabilities of US forces. While conducting these operations, the software will utilize bathymetry, ocean layer, and depth data to show how each of these things impacts the movement, communications, and sensors of that are attached to the subsurface vessels.