GACF: Ground-Aerial Collaborative Framework for Large-Scale Emergency Rescue Scenarios

Large-scale emergency rescue tasks require search and rescue units to conduct autonomous exploration and to be able to respond quickly in complex and dangerous environments. This paper proposes a ground-aerial collaborative framework for search and rescue scenarios. The heterogeneous collaborative system is composed of a single unmanned aerial vehicle(UAV) and a single unmanned ground vehicle(UGV). The UAV is responsible for providing a global topological map of the airspace perspective and performing path planning, and the UGV is responsible for local obstacle avoidance. Collaborative exploration has improved the autonomous exploration capabilities of unmanned systems in large-scale emergency rescue scenarios. This paper proposes a road network extraction and topology construction algorithm based on a airspace perspective image, providing a global traversable map. A path planning algorithm based on the global topological map of airspace is proposed, which improves the efficiency of UGV's exploratory progress. In the real scene experiment, compared with the single UGV with a priori map, the time consumption of GACF is reduced by 59%, which greatly improves the exploration efficiency of the unmanned system in practical applications.