Think before Go: Hierarchical Reasoning for Image-goal Navigation

TL;DR

The paper introduces HRNav, a hierarchical framework for image-goal navigation that combines high-level planning with low-level reinforcement learning to improve navigation efficiency and reduce wandering.

Contribution

It proposes a novel hierarchical reasoning approach that decomposes the task into planning and execution, enhancing performance over existing end-to-end methods.

Findings

HRNav outperforms existing methods in simulation and real-world tests.

The hierarchical approach reduces wandering and improves goal-reaching success.

The Wandering Suppression Penalty effectively minimizes unnecessary exploration.

Abstract

Image-goal navigation steers an agent to a target location specified by an image in unseen environments. Existing methods primarily handle this task by learning an end-to-end navigation policy, which compares the similarities of target and observation images and directly predicts the actions. However, when the target is distant or lies in another room, such methods fail to extract informative visual cues, leading the agent to wander around. Motivated by the human cognitive principle that deliberate, high-level reasoning guides fast, reactive execution in complex tasks, we propose Hierarchical Reasoning Navigation (HRNav), a framework that decomposes image-goal navigation into high-level planning and low-level execution. In high-level planning, a vision-language model is trained on a self-collected dataset to generate a short-horizon plan, such as whether the agent should walk through the door or down the hallway. This downgrades the difficulty of the long-horizon task, making it more amenable to the execution part. In low-level execution, an online reinforcement learning policy is utilized to decide actions conditioned on the short-horizon plan. We also devise a novel Wandering Suppression Penalty (WSP) to further reduce the wandering problem. Together, these components form a hierarchical framework for Image-Goal Navigation. Extensive experiments in both simulation and real-world environments demonstrate the superiority of our method.