GEMINUS: Dual-aware Global and Scene-Adaptive Mixture-of-Experts for End-to-End Autonomous Driving

2025-07-18

AI Summary

To address the challenge of balancing robustness and scene adaptability for end-to-end autonomous driving in complex, dynamic traffic environments, this paper proposes the Dual-Perception Mixture-of-Experts (MoE) framework. The method decouples global driving priors from scene-specific policies and introduces a dual-perception router that jointly models scene features and routing uncertainty, enabling dynamic activation of both global experts and scene-adaptive expert groups to synergistically fuse diverse driving skills. The model operates end-to-end from monocular visual input, with joint training and inference. Evaluated in the Bench2Drive closed-loop benchmark, it achieves state-of-the-art performance, improving driving score, success rate, and multi-capability mean by 7.67%, 22.06%, and 19.41%, respectively.

Technology Category

• Machine Learning: Mixture of Experts (MoE)
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

End-to-end autonomous driving requires adaptive and robust handling of complex and diverse traffic environments. However, prevalent single-mode planning methods attempt to learn an overall policy while struggling to acquire diversified driving skills to handle diverse scenarios. Therefore, this paper proposes GEMINUS, a Mixture-of-Experts end-to-end autonomous driving framework featuring a Global Expert, a Scene-Adaptive Experts Group, and equipped with a Dual-aware Router. Specifically, the Global Expert is trained on the overall dataset, possessing robust performance. The Scene-Adaptive Experts are trained on corresponding scene subsets, achieving adaptive performance. The Dual-aware Router simultaneously considers scenario-level features and routing uncertainty to dynamically activate expert modules. Through the effective coupling of the Global Expert and the Scene-Adaptive Experts Group via the Dual-aware Router, GEMINUS achieves adaptive and robust performance in diverse scenarios. GEMINUS outperforms existing methods in the Bench2Drive closed-loop benchmark and achieves state-of-the-art performance in Driving Score and Success Rate, even with only monocular vision input. Furthermore, ablation studies demonstrate significant improvements over the original single-expert baseline: 7.67% in Driving Score, 22.06% in Success Rate, and 19.41% in MultiAbility-Mean. The code will be available at https://github.com/newbrains1/GEMINUS.

Problem

• Adaptive handling of diverse traffic environments in autonomous driving
• Overcoming limitations of single-mode planning with diverse driving skills
• Dynamic expert module activation for robust performance in varied scenarios

Innovation

• Mixture-of-Experts framework with Global Expert
• Scene-Adaptive Experts Group for diverse scenarios
• Dual-aware Router for dynamic expert activation