The convergence of miniaturisation and advanced on-board compute is enabling a new class of cost-effective, multi-sensor Unmanned Aerial System (UAS) that are reshaping the modern battlespace. However, unlocking their full potential requires a shift from remotely piloted data collectors to truly autonomous, intelligent and coordinated teams. This paper introduces a novel, three-layer cognitive architecture that integrates a utility-based system for strategic prioritisation, a goal oriented action planner for tactical flexibility, and a stigmergic layer for decentralised coordination. Validation in a simulated military Intelligence, Surveillance, and Reconnaissance (ISR), attack and supply mission demonstrates that the stigmergic coordination layer yields a profound increase in efficiency, reducing mission completion time by 46.5% and total distance travelled by 47.7%. The framework´s data-driven design also facilitates the agile development of tactics, allowing new mission sets and behaviours to be rapidly configured. The results validate a practical path towards scalable, intelligent autonomous systems capable of emergent, coordinated action in dynamic environments.
Hemme et al. (Wed,) studied this question.