OmniVLA: Physically-Grounded Multimodal VLA with Unified Multi-Sensor Perception for Robotic Manipulation

TL;DR

OmniVLA introduces a physically-grounded, multi-sensor perception framework for robotic manipulation, significantly enhancing task success rates by integrating infrared, radar, and audio sensors with RGB images.

Contribution

The paper proposes a novel omni-modality VLA model that unifies multiple sensor modalities into a sensor-masked image, enabling efficient, physically-grounded perception for robotic tasks.

Findings

Achieves 84% success rate on real-world tasks.

Outperforms RGB-only models by 59%.

Demonstrates higher learning efficiency and generalization.

Abstract

Vision-language-action (VLA) models have shown strong generalization for robotic action prediction through large-scale vision-language pretraining. However, most existing models rely solely on RGB cameras, limiting their perception and, consequently, manipulation capabilities. We present OmniVLA, an omni-modality VLA model that integrates novel sensing modalities for physically-grounded spatial intelligence beyond RGB perception. The core of our approach is the sensor-masked image, a unified representation that overlays spatially grounded and physically meaningful masks onto the RGB images, derived from sensors including an infrared camera, a mmWave radar, and a microphone array. This image-native unification keeps sensor input close to RGB statistics to facilitate training, provides a uniform interface across sensor hardware, and enables data-efficient learning with lightweight per-sensor projectors. Built on this, we present a multisensory vision-language-action model architecture and train the model based on an RGB-pretrained VLA backbone. We evaluate OmniVLA on challenging real-world tasks where sensor-modality perception guides the robotic manipulation. OmniVLA achieves an average task success rate of 84%, significantly outperforms both RGB-only and raw-sensor-input baseline models by 59% and 28% respectively, meanwhile showing higher learning efficiency and stronger generalization capability.