Senser Fusion and Hierarchical Inspection Strategy for Aerial-Ground Multi-robot System

Abstract

Infrastructure inspection requires balancing coverage and detail for effective assessment.This paper presents a multi-robot inspection system combining aerial and ground platforms through a hierarchical approach.A hexacopter drone performs rapid site mapping, while specialized ground robots (hexapod and tracked variants) conduct detailed inspections.The system implements bidirectional learning where aerial mapping guides ground robot deployment, while ground inspection data refines aerial strategies.Multi-modal sensor data integration uses an Extended Kalman Filter framework to create unified structural health representations.Experimental results demonstrate 96.3% feature extraction accuracy, sub-millimeter registration between sensors, 3.2-second anomaly response time, and 99.7% collision-free operation.The system shows significant potential for inspecting large-scale infrastructure including bridges, industrial facilities, and power plants.