A Risk-Based Geospatial Optimization Framework for UAV Maritime Surveillance Path Planning in the North Natuna Sea

Abstract

Maritime surveillance in strategically critical waters remains challenged by severe operational constraints limited fuel allocation, restricted vessel availability, and constrained aircraft endurance, creating extensive surveillance gaps in designated Areas of Responsibility (AoR). This research develops a risk-based geospatial optimization framework for Unmanned Aerial Vehicle (UAV) maritime surveillance path planning that explicitly addresses these surveillance gaps. The framework integrates four methodological stages: (1) strategic context analysis using PESTLE-SWOT and Analytical Network Process for optimal platform selection, (2) geospatial risk surface modeling using Kernel Density Estimation (KDE) applied to 159 historical maritime violations spanning 2021–2023, (3) multi-objective path optimization balancing threat detection efficiency with operational constraints, and (4) quantitative validation through comparative analysis with baseline routes. KDE analysis reveals highly non-uniform threat distribution across the study area, with significant violation clustering in specific geographic hotspots. Critically, the AoR encompasses extensive Low Risk Zones (83.3% of total area) that represent unmonitored regions where illicit activities possess higher potential to occur undetected due to sparse conventional patrol coverage. These Low-Risk Zones are not inherently safe but rather surveillance gaps requiring persistent monitoring. The optimized UAV route strategically allocates surveillance resources to cover both High-Risk concentration zones and the critical Low Risk areas, achieving 89.3% risk coverage efficiency and 43.7% improvement in threat detection efficiency compared to conventional uniform-coverage baseline routes. The optimized route spans 2,447 km, completed within 13.59 hours of flight time, well within UAV endurance constraints. This research provides a scientifically rigorous framework that integrates geospatial risk modeling with multi-objective optimization, moving beyond conventional geometric path planning toward a data-driven, threat-aware surveillance strategy. The methodology demonstrates that systematic identification and coverage of surveillance gaps particularly in Low-Risk Zones significantly enhances Maritime Domain Awareness and maritime security effectiveness.