Search papers, labs, and topics across Lattice.
This paper introduces a hybrid UAV path planning framework that combines trajectory prediction with a Priority-aware Dynamic Window Approach (P-DWA) to improve adaptability and reduce collision risks in dynamic environments. The framework uses a trajectory prediction model incorporating time weights and uncertainty quantification based on dynamic obstacle positions. By implementing a priority queue mechanism and a risk-aware collision cost function, the approach effectively avoids local optima and improves planning success rate compared to EGOv2 and DP, as validated through simulations and real-world flight tests.
UAVs can nimbly navigate complex, dynamic environments thanks to a new path planning framework that anticipates obstacle trajectories and prioritizes safe routes.
Currently, path planning for unmanned aerial vehicles (UAVs) in dynamic environments still faces risks and challenges such as poor adaptability and high collision risks caused by frequent environmental changes. This paper proposes a hybrid planning framework that integrates trajectory prediction with the Priority-aware Dynamic Window Approach (P-DWA). The framework constructs a trajectory prediction model based on dynamic obstacle position data, integrating time weights and uncertainty quantification. During the path search process, a priority queue mechanism is implemented. This mechanism is combined with a risk-aware collision cost function to avoid local optima. Simulation results demonstrate that the proposed method outperforms EGOv2 and DP in dynamic obstacle scenarios, particularly in terms of planning success rate and obstacle avoidance. Real-world UAV flight tests further validate the method's effectiveness in complex dynamic environments, showcasing its robustness and reliability.
