# HM3D-M3：面向多智能体、多模态、多对象目标导航的基准

> 原文链接：https://robotics.hellobike.com/papers/2026-10-14-2610-20121
> 频道：papers | 更新日期：2026-10-14 | 标签：arxiv, cs.RO
> 核心摘要：HM3D-M3针对多智能体多模态多对象导航，构建基准并以两种基线在室内场景和三机器人团队上验证。

# HM3D-M3：面向多智能体、多模态、多对象目标导航的基准

::: v-pre
> 原文链接：[HM3D-M3: A Benchmark for Multi-Agent, Multi-Modal, Multi-Object Goal Navigation](https://arxiv.org/abs/2610.20121)（arXiv 2610.20121 · 发布日期：2026-10-14）

**HM3D-M3: A Benchmark for Multi-Agent, Multi-Modal, Multi-Object Goal Navigation**

HM3D-M3针对多智能体多模态多对象导航，构建基准并以两种基线在室内场景和三机器人团队上验证。

## 核心要点

- 现有基准分别研究多智能体、多模态和多对象导航，缺少在同一回合中同时覆盖三者的评测。
- HM3D-M3中，K个智能体从分布位置出发，寻找以类别、图像或描述给出的M个无序对象目标。
- 基准提供1000个验证回合、3926个目标、100K个训练样本，并提出目标级SR和归一化SPL指标。
- 作者贡献HiGA和LaViRA-M3两种基线，SR达38--41%，并部署LaViRA-M3到异构三机器人团队。

**验证范围与局限**：验证范围包括照片级真实室内场景中的基准评测，以及LaViRA-M3在异构三机器人团队上的部署。

## 英文摘要 (Abstract)

Rapid progress in AI agents and robotics makes it a matter of the near future that humans are served by a team of robots. Such a team will receive several goals at once, each specified in whatever form is convenient for the human, such as a category name, a photo, or a spoken description. Existing benchmarks study multi-agent, multi-modal, and multi-object goal navigation in isolation, and a benchmark that places all three in one episode is missing. To fill this gap, we introduce HM3D-M3. A team of K agents starts at distributed locations in a photorealistic indoor scene and searches for an unordered set of M object goals, each given as an open-vocabulary category, an instance image, or an instance description. HM3D-M3 provides 1,000 validation episodes over 3,926 goals from 379 open-vocabulary categories, 100K observation-assignment training samples, and two team-level metrics, goal-level SR and makespan-normalized SPL. We further contribute two baselines. HiGA is a trained hierarchical goal assigner on top of a frozen open-vocabulary navigation policy, and LaViRA-M3 is a training-free team of VLM agents with shared mapping and progress analysis. Adapted multi-robot and zero-shot methods reach 16--27% SR, while our baselines reach 38--41% with complementary strengths and distinct failure modes. Furthermore, we deploy LaViRA-M3 on a heterogeneous three-robot team. Project page: https://robo-lavira.github.io/hm3d-m3/

作者：Hongyu Ding, Kanyu Xu, Mingjia Chen, Zixuan Chen, Jieqi Shi, Jing Huo, Yifan Zhang, Yang Gao
:::

## 原始参考来源 (Sources)

- [arXiv 2610.20121：HM3D-M3: A Benchmark for Multi-Agent, Multi-Modal, Multi-Object Goal Navigation](https://arxiv.org/abs/2610.20121)