Self-organising Distributed Sensor Fusion Networks for Hierarchical Swarm Control and Supervision

摘要

Mobile sensor networks can be realised using robot swarms where simple robots interact only locally to achieve swarm scalability and robustness. One of the main challenges is to develop suitable sensor fusion methods, both for autonomous swarms and human supervisory control, without gravely reducing the benefits of decentralisation. Hence, we introduce the Self-organising Hierarchical Extending (SHE) approach for distributed sensor fusion in robot swarms under communication and control constraints found in harsh, heterogeneous, and dynamic environments. The SHE approach is designed for fast autonomous swarm control in local environments, radially extending the sensor-effector range of a central node for global autonomous or human supervisory control. We discuss possible approaches to distributed object search and a SHE fusion architecture. An exemplary simulation based on the virtual pheromone approach provides a proof of concept for a self-organised, hierarchical, and extending fusion topology with a minimal robot controller.