Answers are grounded in this paper and cite their sources.
To address the challenges of coordinating global path planning with local obstacle avoidance and improving terminal convergence after obstacle avoidance in structured environments, this paper proposes an autonomous unmanned aerial vehicle (UAV) navigation framework integrating enhanced A-star (A*) search, vector field histogram plus (VFH+), and goal-directed recovery, implemented on an onboard embedded platform. The framework employs a D435i depth camera as a unified perception source. Visual simultaneous localization and mapping (SLAM) constructs a two-dimensional occupancy grid map for A* global planning, while point-cloud filtering and polar histograms support VFH+ local obstacle avoidance. An asymmetric steering preference is introduced to guide bypass-direction selection, and a goal-directed recovery controller guides the UAV toward the destination after obstacle avoidance. A finite-state machine coordinates path following, obstacle avoidance, goal-directed recovery, and altitude recovery. Simulation results show that the proposed method reduces flight time by approximately 60.0% compared with A* combined with the artificial potential field (APF) method and decreases endpoint error by 25.0% compared with A*+VFH+ without goal-directed recovery, achieving an endpoint error of 0.24 m and a path efficiency of 0.93. Across five repeated physical flight trials, the proposed method achieved a mean endpoint error of 0.38 ± 0.027 m and a path efficiency of 0.920 ± 0.007, demonstrating repeatable navigation performance and the feasibility of onboard implementation under the tested conditions.
No takes yet. Share an insight, caveat, or question.
Yang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: