Project | Amenable

AMENABLE addresses a fundamental challenge in AI-driven environment monitoring: the gap between the informal, task-specific concept hierarchies typical of robotics and sensor systems, and the formal, reusable knowledge structures of knowledge engineering. The project combines three research threads: collaborative 3D semantic scene understanding across heterogeneous robot teams; expert-in-the-loop machine learning that grounds sensor-derived representations in formal ontological knowledge; and explainable AI for resource-constrained edge devices, making smart sensor reasoning interpretable to both engineers and domain experts. Natural language inference is used throughout to maintain consistency as knowledge structures evolve. The methods are validated in autonomous indoor robot exploration and ecological monitoring.