DroneWAM: Efficient World Action Model for Drone Visual Navigation

2026-09-27

AI Summary

This study addresses the challenge of balancing accuracy and efficiency in world models for UAV visual navigation, where explicit image generation incurs prohibitive computational costs. To this end, we propose an efficient world action model that pioneers a JEPA-based architecture to predict future states within a latent representation space, thereby circumventing pixel-level generation. The method incorporates a pretrained Resampler for feature compression and introduces a scene-aware adaptive rollout strategy to dynamically optimize computational allocation. Furthermore, we construct DroneNav-6D, a novel dataset incorporating wind disturbances. Experimental results demonstrate that the proposed model achieves state-of-the-art trajectory accuracy while reducing the average prediction depth from 8 to 4.58, significantly enhancing both inference efficiency and navigation precision.

Abstract

World-action models give visual navigation agents a way to anticipate how candidate actions will change future observations and to act from the predicted consequences. For drones, this capability must operate under tight accuracy and efficiency constraints. We present DroneWAM, an efficient world-action model for drone visual navigation. DroneWAM adopts a JEPA-based architecture to model future states directly in representation space, avoiding the cost of explicit future image generation. A pretrained Resampler further compresses dense encoder features into fewer latent tokens, reducing the computation repeated at each imagined step. We also introduce adaptive rollout, where a preference-trained Gate adaptively allocates prediction depth according to the current scene. To support learning under richer aerial motion, we construct DroneNav-6D, a simulated visual navigation dataset with synchronized RGB observations, 6-DoF flight trajectories, control commands, and randomized wind disturbances. On DroneNav-6D, DroneWAM achieves the best trajectory accuracy among the compared methods. Adaptive rollout further reduces the average prediction depth from 8 to 4.58 while improving trajectory accuracy, demonstrating that predictive computation can be allocated more effectively across scenes. \href{https://github.com/1e12Leon/DroneWAM}{Codes and data} will be released.

Problem

• drone visual navigation
• world-action model
• computational efficiency
• 6-DoF flight trajectories

Innovation

• World Action Model
• JEPA
• Adaptive Rollout
• Visual Navigation
• DroneNav-6D