Deep Terrain Shaping: Coordinated Effects and Maneuver Using Swarms

In SBIR A244-003 (Deep Terrain Shaping), the Army Applications Lab (AAL) seeks to revolutionize the US Army's terrain shaping process, transitioning from traditional methods to advanced autonomy, AI, robotics, and sensor technology. Novel Engineering, Inc. proposes Coalesce, an autonomous Multi-Agent Command and Control (C2) system, designed to manage terrain shaping operations and includes a collaborative UxV system for detecting and engaging hostile threats. Coalesce is a platform-agnostic system that equips any autonomous vehicle with the ability to collaboratively complete complex objectives. CoalesceÆs core components include a mission planning and monitoring system, a distributed Common Operating Picture (COP), and a custom navigation stack. The system is fully distributed, every platform has all the components needed to reason over their objectives, their swarm-mates current tasking, and the current situation to decide on their next best action even in comms degraded or denied environments. This is controlled from a custom User Interface allowing an operator to plan, monitor, and update swarm activities enabling the swarm to operate at machine speed with human-in-the-loop approving all lethal action. Envisioned to function as the C2 for a comprehensive system of systems, Coalesce operates across extensive battlefield areas, up to 150km ahead of the Forward Line of Troops (FLOT). It utilizes long-range sensors to detect movement and threats. Large carrier Unmanned Ground Vehicles (UGVs) deploy small UAS and UGVs deep into the battlespace upon threat identification. These deployed units, equipped with lethal and non-lethal effectors, execute tasks including detection, identification, engagement, and threat assessment. Ā To enhance Coalesce's capabilities, we propose the following significant upgrades to its COP as well as the creation of new swarm tactics to carry out disrupt and fix terrain shaping operations: A situation module that classifies entities in the COP by hierarchical types and provides additional information like expected movement and capability characteristics. A synchronization module that efficiently compares COPs between swarm agents and updates parts of one or both that are out of date. A distributed track fusion algorithm based on the latest particle filter research to enable more accurate tracks built from multiple sensor sources. New swarm tactics to complete all parts of the terrain shaping operation In Phase 1, our approach includes refining system design by collaborating with the user community to derive mission requirements. This phase will detail the selection of vehicles, sensors, tactics, and algorithms, validated simulations or live demonstrations as appropriate. To do this, we are partnering with L3Harris Technologies, Inc for mentorship and SME support in our process and validation of our proposed approaches. The findings and developments will be comprehensively reported at the end of Phase 1.