DPNet: Efficient Dead-End Prediction and Avoidance for Vision-Based UAV Navigation
2026-08-17 • Robotics
Robotics
AI summaryⓘ
The authors developed a method to help drones avoid getting stuck in dead ends during flight using cameras and depth sensors. They created a small neural network that predicts how far and in what direction a dead end is within the drone's view. This prediction helps the drone choose safe paths ahead of time without slowing down. Their system works well in simulations and real-world tests without needing extra training on new data. It also replans routes very quickly to keep the drone flying smoothly.
Unmanned Aerial Vehiclesdead-end predictionRGB-D inputneural networktrajectory planningnavigationreplanning frequencysimulation benchmarksreal-world transferfield of view
Authors
Ruibin Zhang, Lun Pan, Zelong Xia, Jialiang Hou, Fei Gao
Abstract
Vision-based Unmanned Aerial Vehicles (UAVs) often suffer from navigation failures in dead ends due to limited sensing accuracy and range. To address this challenge, this paper proposes a systematic solution for efficient dead-end prediction and avoidance. The proposed method introduces a lightweight neural network to predict the relative distance and bearing of potential dead ends within the current field of view using RGB-D inputs. These predictions prune a predefined, compact trajectory library, enabling the planner to proactively avoid dead ends while maintaining navigational smoothness. Notably, our approach transfers across real-world scenarios without manual annotation or fine-tuning on real-world data. The system achieves high-frequency replanning at 50 Hz onboard. Extensive simulation benchmarks demonstrate superior performance in success rate, flight time, and trajectory length, and real-world experiments further validate its effectiveness in complex scenarios.