Geometry-Aware Moving Viewpoint Particle Swarm Optimisation for Efficient UAV Facade Coverage

Abstract

Inspection of high-rise building façades is critical for structural assessment, emergency response, and risk mitigation in dense urban environments. While Unmanned Aerial Vehicles (UAVs) provide a safe and flexible platform for data acquisition, the effectiveness of UAVbased inspection is fundamentally constrained by coverage path planning. Conventional optimisation approaches typically operate over static, precomputed viewpoint sets or abstract search spaces, limiting adaptability in complex, concave, and partially occluded geometries. This work proposes a geometry-aware moving-viewpoint Particle Swarm Optimisation (VPSO) framework for UAV façade coverage planning. In the proposed formulation, viewpoints are treated as movable agents constrained by local surface geometry, and the search space is reduced from three spatial degrees of freedom to a single angular parameter. By embedding geometric feasibility directly into the optimisation process, the method integrates coverage optimisation with continuous surface-aware motion. The approach is evaluated against a conventional PSO baseline across three building geometries of increasing complexity. Results demonstrate improved optimisation stability, higher coverage efficiency, and enhanced robustness in non-convex and occluded environments. The proposed framework provides a foundation for unified coverage and trajectory optimisation in UAV-based structural inspection.