A UAV Exploration Planning Method Based on Improved Bio-Inspired Neural Network
Sheng Ma, Yu Zhou, Zhifa Gao, Yuzhe Li, Qinghua Hao, Ling Zhao · 2025
This paper presents an UAV exploration planning method based on improved 3D-bio-inspired neural networks (3D-BNN) to address the challenges of inefficient path planning and susceptibility to local optima in autonomous drone exploration of unknown environments. The proposed method integrates frontier boundary regions with 3D-BNN, enabling global guidance for UAV exploration. Furthermore, a hierarchical exploration structure is developed, combining Frontier boundary region, 3D-BNN and viewpoints to synergize global path planning with local optimization. This dual-layered framework effectively mitigates local optima traps and minimizes redundant exploration. Extensive simulations in the Gazebo environment demonstrate the superiority of the proposed method over the FUEL algorithm, achieving a$10.8 \%-13.9 \%$improvement in exploration speed while enabling faster and more efficient autonomous navigation in unknown spaces.