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

Hugging Face Daily Papers Papers
benchmark self-awareness spatial-cognition embodied-intelligence uav multimodal-llm motion-awareness

Summary

This paper introduces SIS-Bench, a benchmark for evaluating self-awareness and spatial cognition in UAV embodied intelligence using multimodal large language models, and explores motion-aware representations to improve performance.

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.
Original Article
View Cached Full Text

Cached at: 07/16/26, 05:44 PM

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

Source: https://huggingface.co/papers/2607.12477

Abstract

AutonomousUAVsystemsincreasinglyrelyonmultimodallargelanguagemodels(MLLMs)tooperateincomplexreal-worldenvironments.Suchembodiedscenariosrequirenotonlyunderstandingthesurroundingspacebutalsomaintainingacoherentrepresentationoftheagentitself.However,existingUAV-orientedapproachesandbenchmarksremainlargelyenvironment-centric,primarilyfocusingonspatialunderstandingtasks,withtheagent’sself-awarenessremainingimplicit.Toaddressthisgap,weintroduceSIS-Bench,abenchmarkforevaluatingembodiedspatialintelligenceinUAVscenariosunderaunifiedself-in-spaceformulation.SIS-Benchorganizesevaluationalongtwocomplementarydimensions,spaceandself,andathree-levelhierarchyofperception,memory,andreasoning.Itcontains4,856question--answerpairsacross13tasksderivedfrom1,646real-worldUAVvideosthroughatask-conditionedconstructionpipelinewithexpertverification.ExtensiveevaluationsrevealthatcurrentMLLMsexhibitfundamentallimitationsinmodelingdynamicandagent-centeredprocesses.Inparticular,weobserveaclearimbalancebetweenspatialcognitionandself-awareness,aswellasaprogressiveperformancedegradationacrosscognitivelevels.Motivatedbythesefindings,wefurtherexploreamotion-awarerepresentationthatincorporatesself-relateddynamicsthroughopticalflowandvisualfeaturefusion.Experimentalresultsshowthatmodelingagentmotionconsistentlyimprovesperceptionandmemoryperformance,notonlyinspatialcognitionbutalsoinself-awareness,andgeneralizestodownstreamUAVdecision-makingtasks.Ourresultshighlighttheimportanceofself-awarenessforadvancingembodiedspatialintelligence,andprovidebothanewbenchmarkandempiricalevidenceformotion-awareself-in-spacemodeling.

View arXiv pageView PDFProject pageGitHub4Add to collection

Get this paper in your agent:

hf papers read 2607\.12477

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper1

#### choucsan/SIS-Motion Updated1 day ago • 88 • 2

Datasets citing this paper3

#### choucsan/SIS-Bench Viewer• Updatedabout 10 hours ago • 4.86k • 591 • 2 #### choucsan/OpenUAV-QA Preview• Updated1 day ago • 372 • 2 #### choucsan/SIS-Motion-54K Viewer• Updated1 day ago • 54.3k • 167 • 4

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2607.12477 in a Space README.md to link it from this page.

Collections including this paper1

Similar Articles

ESI-Bench: Towards Embodied Spatial Intelligence that Closes the Perception-Action Loop

Hugging Face Daily Papers

Introduces ESI-BENCH, a comprehensive benchmark for embodied spatial intelligence built on OmniGibson, covering 10 task categories and 29 subcategories. Experiments show active exploration substantially outperforms passive approaches, with failures mainly due to action blindness rather than perception, revealing a metacognitive gap in models compared to humans.

OVO-S-Bench: A Hierarchical Benchmark for Streaming Spatial Intelligence in Multimodal LLMs

Papers with Code Trending

OVO-S-Bench introduces a comprehensive human-annotated benchmark of 1,680 questions across 348 videos to evaluate streaming spatial intelligence in multimodal LLMs, revealing that even the best model (Gemini-3.1-Pro) trails human experts by 27 points. The benchmark exposes key limitations including allocentric mapping as a major bottleneck and chain-of-thought reasoning amplifying spatial errors.