Answers are grounded in this paper and cite their sources.
Low-altitude UAV (Unmanned Aerial Vehicle) path planning in urban areas must balance energy efficiency, safety, terrain clearance, climb constraints, and restricted-airspace avoidance. To address this problem, this study formulates urban low-altitude UAV path planning as a constrained bi-objective optimization model that simultaneously minimizes flight energy consumption and cumulative threat exposure. An enhanced multi-objective Grey Wolf Optimizer, denoted as IMOGWO, is developed to solve the proposed energy–threat-aware path-planning model. To improve search performance in fragmented urban feasible spaces, two path-oriented candidate-generation mechanisms are incorporated. The first is an elite-guided Gaussian multiplicative perturbation, which refines waypoint offsets around high-quality feasible paths. The second is a center-constrained quasi-opposite candidate generation, which explores complementary bypass corridors by sampling between the classical opposite point and the search-space center, thereby reducing the risk of generating extreme infeasible solutions. Experiments are conducted on two representative urban low-altitude scenarios: Map 1 with 100 buildings and one no-fly zone and Map 2 with 150 buildings. An additional Map 2 configuration with dense buildings and three no-fly zones is also evaluated as Scenario 3 together with a non-GWO NSGA-II baseline. The proposed IMOGWO achieves the highest hypervolume (HV) values in both original scenarios. In the additional Scenario 3, the 30-run mean HV of IMOGWO is also higher than those of the four baselines, including NSGA-II. Compared with MOPSO, MOGWO, and MOGGO, IMOGWO improves HV by 19.05%, 2.36%, and 3.46% in Map 1 and by 26.51%, 12.17%, and 4.22% in Map 2, respectively. The results demonstrate that the proposed method improves Pareto-set quality and provides more practical energy–threat trade-off paths for UAV operations in constrained urban low-altitude airspace.
No takes yet. Share an insight, caveat, or question.
Zhu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: