OmniGen: Unified Multimodal Sensor Generation for Autonomous Driving

TL;DR

OmniGen introduces a unified framework for generating aligned multimodal sensor data in autonomous driving, improving efficiency and consistency across LiDAR and camera modalities using a shared BEV space and advanced reconstruction techniques.

Contribution

The paper presents OmniGen, a novel multimodal sensor generation model that unifies LiDAR and camera data in a shared BEV space with a new reconstruction method and controllable generation capabilities.

Findings

Achieves high-quality aligned multimodal sensor data synthesis.

Demonstrates improved multimodal consistency and flexibility.

Outperforms existing single-modality generative approaches.

Abstract

Autonomous driving has seen remarkable advancements, largely driven by extensive real-world data collection. However, acquiring diverse and corner-case data remains costly and inefficient. Generative models have emerged as a promising solution by synthesizing realistic sensor data. However, existing approaches primarily focus on single-modality generation, leading to inefficiencies and misalignment in multimodal sensor data. To address these challenges, we propose OminiGen, which generates aligned multimodal sensor data in a unified framework. Our approach leverages a shared Bird\u2019s Eye View (BEV) space to unify multimodal features and designs a novel generalizable multimodal reconstruction method, UAE, to jointly decode LiDAR and multi-view camera data. UAE achieves multimodal sensor decoding through volume rendering, enabling accurate and flexible reconstruction. Furthermore, we incorporate a Diffusion Transformer (DiT) with a ControlNet branch to enable controllable multimodal sensor generation. Our comprehensive experiments demonstrate that OminiGen achieves desired performances in unified multimodal sensor data generation with multimodal consistency and flexible sensor adjustments.