Integration of Mixture of Experts and Multimodal Generative AI in Internet of Vehicles: A Survey

TL;DR

This survey addresses enabling Artificial General Intelligence in the Internet of Vehicles by integrating Mixture-of-Experts (MoE) with multimodal Generative AI (GAI). It covers MoE architectures—sparse gating, distributed execution, all-to-all communications, load balancing—and their applicability to IoV tasks such as intelligent traffic management and autonomous driving, alongside GAI modalities including GANs, VAEs, diffusion models, and large language/vision models. The paper analyzes distributed perception, cooperative decision-making, and generative modeling for simulation, detailing how MoE can allocate specialized experts to different regions or modalities while maintaining real-time performance. It identifies privacy-preserving collaboration, energy efficiency, parameter-efficient fine-tuning, and retrieval-augmented generation as critical directions to realize scalable, robust, and safe MoE-GAI IoV systems with broad practical impact on autonomous transportation and smart cities.

Abstract

Generative AI (GAI) can enhance the cognitive, reasoning, and planning capabilities of intelligent modules in the Internet of Vehicles (IoV) by synthesizing augmented datasets, completing sensor data, and making sequential decisions. In addition, the mixture of experts (MoE) can enable the distributed and collaborative execution of AI models without performance degradation between connected vehicles. In this survey, we explore the integration of MoE and GAI to enable Artificial General Intelligence in IoV, which can enable the realization of full autonomy for IoV with minimal human supervision and applicability in a wide range of mobility scenarios, including environment monitoring, traffic management, and autonomous driving. In particular, we present the fundamentals of GAI, MoE, and their interplay applications in IoV. Furthermore, we discuss the potential integration of MoE and GAI in IoV, including distributed perception and monitoring, collaborative decision-making and planning, and generative modeling and simulation. Finally, we present several potential research directions for facilitating the integration.

Figures (6)

• Figure 1: The outline of this survey.
• Figure 2: Internet of Vehicles supported by different generative AI technologies.
• Figure 3: Mixture-of-experts (MoE) Architecture and its applications in IoV.
• Figure 4: The workflow of traffic management leveraging the MoE architecture for future traffic prediction.
• Figure 5: SafePathNet can predict future trajectories and road agents using MoE architecture based on the input scene.