Problem
Research questions and friction points this paper is trying to address.
Vision-Based UAV Navigation
Dead-End Prediction
Dead-End Avoidance
Navigation Failure
DPNet: Efficient Dead-End Prediction and Avoidance for Vision-Based UAV Navigation
🤖 AI Summary
This study addresses navigation failures in visual UAVs operating within dead-end environments by proposing a lightweight network-based trajectory library pruning and high-frequency replanning framework. Relying solely on RGB-D inputs, the method achieves zero-shot cross-scenario transfer without real-world annotation while ensuring both collision avoidance safety and trajectory smoothness. Simulation and real-flight experiments demonstrate that the system enables 50 Hz onboard real-time planning, significantly improving navigation success rates and efficiency in complex scenarios. Consequently, this approach effectively resolves dynamic obstacle avoidance challenges under perception-constrained conditions, offering a robust solution for autonomous flight in confined spaces where traditional methods typically fail.
📝 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.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Dead-End Prediction
Lightweight Neural Network
Trajectory Library Pruning
Zero-shot Transfer
Onboard Replanning
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
No related jobs found.
