# ROMA: LLM System for Real-World Object-Centric Multi-Sensory Active Perception

arXiv: https://arxiv.org/abs/2610.06955 · HTML: https://mlsanity.com/paper/2610.06955/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 48 of its 56 references that are in the map.
- First posted on arXiv 3 Oct 2026.

## Builds on (most cited first, 30 of 48)

- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [Qwen2.5-Omni Technical Report](https://mlsanity.com/paper/2503.20215/) (2025)
- [SpatialVLM: Endowing Vision-Language Models with Spatial Reasoning Capabilities](https://mlsanity.com/paper/2401.12168/) (2024)
- [Qwen3-Omni Technical Report](https://mlsanity.com/paper/2509.17765/) (2025)
- [Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots](https://mlsanity.com/paper/2402.10329/) (2024)
- [AnyGrasp: Robust and Efficient Grasp Perception in Spatial and Temporal Domains](https://mlsanity.com/paper/2212.08333/) (2022)
- [Qwen3.5-Omni Technical Report](https://mlsanity.com/paper/2604.15804/) (2026)
- [Physically Grounded Vision-Language Models for Robotic Manipulation](https://mlsanity.com/paper/2309.02561/) (2023)
- [ForceVLA: Enhancing VLA Models with a Force-aware MoE for Contact-rich Manipulation](https://mlsanity.com/paper/2505.22159/) (2025)
- [Touch and Go: Learning from Human-Collected Vision and Touch](https://mlsanity.com/paper/2211.12498/) (2022)
- [ObjectFolder 2.0: A Multisensory Object Dataset for Sim2Real Transfer](https://mlsanity.com/paper/2204.02389/) (2022)
- [A Touch, Vision, and Language Dataset for Multimodal Alignment](https://mlsanity.com/paper/2402.13232/) (2024)
- [See, Hear, and Feel: Smart Sensory Fusion for Robotic Manipulation](https://mlsanity.com/paper/2212.03858/) (2022)
- [Sparsh: Self-supervised touch representations for vision-based tactile sensing](https://mlsanity.com/paper/2410.24090/) (2024)
- [SEAL: Self-supervised Embodied Active Learning using Exploration and 3D Consistency](https://mlsanity.com/paper/2112.01001/) (2021)
- [Octopi: Object Property Reasoning with Large Tactile-Language Models](https://mlsanity.com/paper/2405.02794/) (2024)
- [AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors](https://mlsanity.com/paper/2502.12191/) (2025)
- [Beyond Sight: Finetuning Generalist Robot Policies with Heterogeneous Sensors via Language Grounding](https://mlsanity.com/paper/2501.04693/) (2025)
- [MultiPLY: A Multisensory Object-Centric Embodied Large Language Model in 3D World](https://mlsanity.com/paper/2401.08577/) (2024)
- [Vision in Action: Learning Active Perception from Human Demonstrations](https://mlsanity.com/paper/2506.15666/) (2025)
- [FoAR: Force-Aware Reactive Policy for Contact-Rich Robotic Manipulation](https://mlsanity.com/paper/2411.15753/) (2024)
- [ManiWAV: Learning Robot Manipulation from In-the-Wild Audio-Visual Data](https://mlsanity.com/paper/2406.19464/) (2024)
- [ESI-Bench: Towards Embodied Spatial Intelligence that Closes the Perception-Action Loop](https://mlsanity.com/paper/2605.18746/) (2026)
- [MLA: A Multisensory Language-Action Model for Multimodal Understanding and Forecasting in Robotic Manipulation](https://mlsanity.com/paper/2509.26642/) (2025)
- [Play to the Score: Stage-Guided Dynamic Multi-Sensory Fusion for Robotic Manipulation](https://mlsanity.com/paper/2408.01366/) (2024)
- [Digitizing Touch with an Artificial Multimodal Fingertip](https://mlsanity.com/paper/2411.02479/) (2024)
- [PACS: A Dataset for Physical Audiovisual CommonSense Reasoning](https://mlsanity.com/paper/2203.11130/) (2022)
- [ActiveVLA: Injecting Active Perception into Vision-Language-Action Models for Precise 3D Robotic Manipulation](https://mlsanity.com/paper/2601.08325/) (2026)

## Abstract (arXiv)

Humans inherently understand the physical world through an active process. When sensory evidence is insufficient to infer physical properties, we naturally interact with the environment by deciding what information is missing, how to acquire it, and when sufficient evidence has been obtained. In stark contrast, existing multi-sensory robot systems mainly integrate sensory inputs rather than actively acquiring missing evidence through interactions. In this work, we introduce ROMA, an LLM-based system for Real-World Object-Centric Multi-Sensory Active Perception. ROMA integrates vision, audio, tactile, and force sensing into a reasoning-interaction-feedback loop. The model identifies missing evidence and determines the target objects, interactions, and modalities, while a physical interface executes the selected interactions and collects the multi-sensory feedback. To support this capability, we construct ROMI-2K, a large-scale real-world multi-sensory object interaction dataset covering nearly 2,000 objects and 6 atomic interactions with synchronized sensory feedback. Building on these data, we develop a two-stage training framework that aligns sensory modalities and equips the LLM to assess evidence sufficiency, select informative interactions, and reason over the multi-sensory feedback. We further characterize active perception as perception chains, where acquired evidence guides subsequent interactions and reasoning, and establish ROMA Bench to evaluate single-attribute, long-horizon multi-attribute, and intent-driven active perception. Experiments show that ROMA can actively acquire missing evidence and solve complex, long-chain multi-sensory perception tasks that existing methods struggle to handle, laying a strong perceptual foundation for active multi-sensory embodied agents.