arXiv 2508.014152026-03-24cs.ROcs.AI

RoboMemory：一种脑启发的多记忆代理框架，用于物理具身系统中的交互环境学习

RoboMemory: A Brain-inspired Multi-memory Agentic Framework for Interactive Environmental Learning in Physical Embodied Systems

FNii-Shenzhen（FNii-深圳）；SSE, CUHK-Shenzhen（SSE，CUHK-深圳）；The Chinese University of Hong Kong, Shenzhen（香港中文大学（深圳））；The University of Hong Kong（香港大学）；National University of Singapore（新加坡国立大学）；The Chinese University of Hong Kong（香港中文大学）；Ising AI

Mingcong Lei, Honghao Cai, Yuyuan Yang, Yimou Wu, Jinke Ren, Zezhou Cui, Liangchen Tan, Junkun Hong, Gehan Hu, Shuangyu Zhu, Shaohan Jiang, Ge Wang, Junyuan Tan… 展开作者

Mingcong Lei, Honghao Cai, Yuyuan Yang, Yimou Wu, Jinke Ren, Zezhou Cui, Liangchen Tan, Junkun Hong, Gehan Hu, Shuangyu Zhu, Shaohan Jiang, Ge Wang, Junyuan Tan, Zhenglin Wan, Zheng Li, Zhen Li, Shuguang Cui, Yiming Zhao, Yatong Han

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出RoboMemory框架，通过整合空间、时间、事件和语义记忆，提升机器人在复杂环境中的长期规划与交互学习能力，实验表明其在EmbodiedBench上成功率提升26.5%，超越现有SOTA模型。

摘要：

具身智能旨在使机器人能够学习、推理并在复杂现实环境中稳健泛化。然而，现有方法常面临部分可观测性、碎片化空间推理和异构记忆整合效率低的问题，限制了长期适应能力。为此，我们引入RoboMemory，一种受大脑启发的框架，通过并行架构统一空间、时间、事件和语义记忆，实现高效的长期规划和交互学习。其核心创新是动态空间知识图谱，用于可扩展且一致的记忆更新，以及带有批评模块的闭环规划器，用于自适应决策。在EmbodiedBench上的大量实验表明，当使用Qwen2.5-VL-72B-Ins实例化时，RoboMemory在平均成功率上比其强基线提高了26.5%，甚至超越了闭源SOTA模型Claude-3.5-Sonnet。现实世界试验进一步证实了其累积学习能力，性能在重复任务中持续提升。我们的结果将RoboMemory定位为一种可扩展的记忆增强具身代理的基础，将认知神经科学的见解与实际机器人自主性结合。

英文摘要：

Embodied intelligence aims to enable robots to learn, reason, and generalize robustly across complex real-world environments. However, existing approaches often struggle with partial observability, fragmented spatial reasoning, and inefficient integration of heterogeneous memories, limiting their capacity for long-horizon adaptation. To address this, we introduce RoboMemory, a brain-inspired framework that unifies Spatial, Temporal, Episodic, and Semantic memory within a parallelized architecture for efficient long-horizon planning and interactive learning. Its core innovations are a dynamic spatial knowledge graph for scalable, consistent memory updates and a closed-loop planner with a critic module for adaptive decision-making. Extensive experiments on EmbodiedBench show that RoboMemory, instantiated with Qwen2.5-VL-72B-Ins, improves the average success rate by 26.5% over its strong baseline and even surpasses the closed-source SOTA, Claude-3.5-Sonnet. Real-world trials further confirm its capability for cumulative learning, with performance consistently improving over repeated tasks. Our results position RoboMemory as a scalable foundation for memory-augmented embodied agents, bridging insights from cognitive neuroscience with practical robotic autonomy.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。