PathPainter: Transferring the Generalization Ability of Image Generation Models to Embodied Navigation
For embodied navigation, this work demonstrates a method to transfer the generalization ability of image generation models to navigation tasks, enabling robots to leverage foundation models for long-range navigation.
The paper proposes a navigation system that uses bird's-eye-view images as global priors and an image generation model to interpret natural language commands and generate traversability masks. The system enables a UAV to complete a 160-meter outdoor long-range navigation task using only a conventional local motion planner.
Bird's-eye-view (BEV) images have been widely demonstrated to provide valuable prior information for navigation. Given the global information provided by such views, two key challenges remain: how to fully exploit this information and how to reliably use it during execution. In this paper, we propose a navigation system that uses BEV images as global priors and is designed for ground and near-ground robotic platforms. The system employs an image generation model to interpret human intent from natural language, identify the target destination, and generate traversability masks. During execution, we introduce cross-view localization to align the robot's odometry with the BEV map and mitigate long-term drift in conventional odometry. We conduct extensive benchmark experiments to evaluate the proposed method and further validate it on a UAV platform. Using only a conventional local motion planner, the UAV successfully completes a 160-meter outdoor long-range navigation task. This work demonstrates how the world-understanding capabilities of foundation models can be transferred to embodied navigation, enabling robots to benefit from the strong generalization ability of existing image generation models.