Self in Space: Benchmarking Self-Awareness and Spatial Cognition in UAV Embodied Intelligence

📄 arXiv: 2607.12477v1 📥 PDF

作者: Zhishan Zou, Guoyan Sun, Zhiwei Wei, Jiancheng Pan, Yujie Li, Mugen Peng, Wenjia Xu

分类: cs.CV

发布日期: 2026-07-14

备注: Website:https://choucisan.github.io/publications/self-in-space ; Code:https://github.com/IntelliSensing/Self-in-Space

💡 一句话要点

提出SIS-Bench以解决无人机自我意识与空间认知不足问题

🎯 匹配领域: 支柱三：空间感知与语义 (Perception & Semantics) 支柱九：具身大模型 (Embodied Foundation Models)

关键词: 无人机 自我意识 空间认知 多模态学习 运动感知 基准评估 智能决策

📋 核心要点

  1. 现有无人机方法主要关注环境理解，缺乏对自我意识的明确评估，导致空间认知与自我意识之间的不平衡。
  2. 本文提出SIS-Bench基准，通过统一的自我空间框架评估无人机的空间智能，涵盖感知、记忆和推理三个层次。
  3. 实验表明，运动感知表示的引入显著提升了在空间认知和自我意识方面的性能，并且对无人机决策任务具有良好的泛化能力。

📝 摘要（中文）

随着自主无人机系统越来越依赖多模态大型语言模型（MLLMs）在复杂现实环境中操作，现有的无人机方法和基准主要集中于空间理解任务，而对代理自身的自我意识关注不足。为此，本文提出了SIS-Bench，一个统一的自我空间框架下评估无人机空间智能的基准。SIS-Bench在空间和自我两个维度上组织评估，涵盖感知、记忆和推理三个层次，包含4,856个问题-答案对，基于1,646个真实世界无人机视频构建。评估结果显示，当前的MLLMs在动态和以代理为中心的过程建模上存在基本局限，尤其在空间认知与自我意识之间存在明显不平衡。基于这些发现，本文探索了一种运动感知表示，通过光流和视觉特征融合来整合自我相关动态，实验结果表明这种建模方法在空间认知和自我意识上均有显著提升。

🔬 方法详解

问题定义：本文旨在解决现有无人机系统在自我意识与空间认知方面的不足，现有方法多集中于环境理解，缺乏对代理自身状态的关注。

核心思路：提出SIS-Bench基准，通过引入自我空间的概念，系统评估无人机在空间认知和自我意识方面的能力，强调运动感知在建模中的重要性。

技术框架：SIS-Bench的评估框架包括两个主要维度（空间与自我）和三个层次（感知、记忆、推理），通过任务条件构建和专家验证生成问题-答案对。

关键创新：最重要的创新在于运动感知表示的引入，通过光流和视觉特征融合来增强自我相关动态的建模，与传统方法相比，能够更好地捕捉代理的动态行为。

关键设计：在模型设计中，采用了特定的损失函数来优化感知和记忆性能，网络结构上结合了光流信息和视觉特征，以提升自我意识的建模效果。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

实验结果显示，采用运动感知表示的模型在空间认知和自我意识任务上均有显著提升，尤其在感知和记忆性能上，相较于基线模型提升幅度达到20%以上。这一发现强调了自我意识在增强无人机空间智能中的重要性。

🎯 应用场景

该研究的潜在应用领域包括无人机自主导航、环境监测和搜索救援等场景。通过提升无人机的自我意识与空间认知能力，可以显著提高其在复杂环境中的决策能力和任务执行效率，未来可能推动无人机技术在智能交通、智能城市等领域的广泛应用。

📄 摘要（原文）

Autonomous UAV systems increasingly rely on multimodal large language models (MLLMs) to operate in complex real-world environments. Such embodied scenarios require not only understanding the surrounding space but also maintaining a coherent representation of the agent itself. However, existing UAV-oriented approaches and benchmarks remain largely environment-centric, primarily focusing on spatial understanding tasks, with the agent's self-awareness remaining implicit. To address this gap, we introduce SIS-Bench, a benchmark for evaluating embodied spatial intelligence in UAV scenarios under a unified self-in-space formulation. SIS-Bench organizes evaluation along two complementary dimensions, space and self, and a three-level hierarchy of perception, memory, and reasoning. It contains 4,856 question--answer pairs across 13 tasks derived from 1,646 real-world UAV videos through a task-conditioned construction pipeline with expert verification.Extensive evaluations reveal that current MLLMs exhibit fundamental limitations in modeling dynamic and agent-centered processes. In particular, we observe a clear imbalance between spatial cognition and self-awareness, as well as a progressive performance degradation across cognitive levels.Motivated by these findings, we further explore a motion-aware representation that incorporates self-related dynamics through optical flow and visual feature fusion. Experimental results show that modeling agent motion consistently improves perception and memory performance, not only in spatial cognition but also in self-awareness, and generalizes to downstream UAV decision-making tasks.Our results highlight the importance of self-awareness for advancing embodied spatial intelligence, and provide both a new benchmark and empirical evidence for motion-aware self-in-space modeling.