外观
论文arXiv 2608.184701 个月前
DevGRU：使用碰撞感知循环模型的深度引导视觉导航
原文链接：DevGRU: Depth-Guided Visual Navigation Using a Collision-Aware Recurrent Model（arXiv 2608.18470 · 发布日期：2026-09-02）
DevGRU: Depth-Guided Visual Navigation Using a Collision-Aware Recurrent Model
DevGRU面向复杂室内导航碰撞问题，结合深度图、点目标与碰撞感知循环模型，在九个场景中验证性能。
核心要点
- 现有视觉导航模型在复杂室内、结构化布局和狭窄通道中易发生碰撞。
- DevGRU以深度图和点目标为条件，通过动作预测器生成碰撞感知未来轨迹。
- 结合碰撞预测器补偿目标位姿估计误差，并主动缓解未来轨迹偏差。
- 在九个场景对比多种方法，显著优于ViNT和NoMaD，且推理时间最快。
验证范围与局限：摘要仅说明在九个不同场景中评估，未说明仿真或真机平台。
英文摘要 (Abstract)
Existing visual navigation models often aim to develop foundation models that can generalize robot navigation across diverse platforms. However, many of these models are prone to collisions when deployed in complex indoor environments, particularly in structured layouts and narrow passages. To address this problem, we propose a depth image- and point-goal-conditioned navigation system, DevGRU. The proposed system employs an action predictor (AP) that generates collision-aware future trajectories, enabling effective avoidance of immediate obstacles. In conjunction with a collision predictor, the AP further compensates for errors accumulated in the goal pose estimation and proactively mitigates future deviations. To evaluate our method, we conducted experiments across nine different scenes and three state-of-the-art approaches - ViNT, NoMaD, and NavDP - as well as four additional variants of ViNT and NoMaD. In terms of navigation performance, DevGRU significantly outperforms ViNT and NoMaD by a large margin. In addition, the proposed model has a relatively small number of trainable parameters, resulting in the fastest inference time among the baselines, particularly outperforming NavDP by 7x in model size and 17x in inference time.
作者：Kyung Min Han, Eunsom Kim, Young J. Kim
来源
本页事实对应的公开页面，以原文为准。