UAV path planning using NSGA-II and BOA algorithms for surveillance applications based on bank-to-turn strategy

Abstract

This study presents a supervisory multi-objective geometric path-planning framework for fixed-wing UAV navigation in complex three-dimensional terrain. The path is represented by three-dimensional waypoints, which constitute the optimization variables. For each candidate path generated by the BTT-enhanced Butterfly Optimization Algorithm (NBOA) or the constraint-aware NSGA-II, BTT-derived maneuver constraints are evaluated within the corresponding optimization loop. Bank-angle, turn-rate, climb/descent, load-factor, terrain-clearance, and related constraint violations therefore influence fitness evaluation, constraint handling, and candidate selection. The method optimizes geometric paths rather than time-parameterized trajectories. Three objectives are considered: path length, energy consumption, and geometric path smoothness. The energy objective incorporates climb-related altitude variation and an SFC-based fuel-consumption proxy, whereas collision avoidance is treated as a feasibility constraint. The framework contains NBOA and NSGA-II as complementary optimization branches. In the comparative experiments reported in this study, the two optimizers are evaluated independently under common planning conditions; the supervisory score provides a generalized scenario-level selection criterion rather than a runtime switching mechanism. Experiments on synthetic and real digital elevation models show that the BTT-enhanced BOA reduces computation time by up to 80% and reduces the converged path length by up to 50% relative to the standard BOA under the tested configurations. Independently evaluated NSGA-II solutions provide feasible Pareto trade-offs among path length, energy consumption, and geometric path smoothness while maintaining the adopted maneuver-feasibility constraints.