Multimodal collaborative perception and digital twin for intelligent resource management and reliable communication in 6G vehicular networks

Abstract

The evolution towards sixth-generation (6G) vehicular-to-everything (V2X) communication presents unprecedented opportunities and significant challenges, particularly in ultra-reliable low-latency communications (URLLC), channel state prediction, blockage management, and intelligent radio resource management (RRM). Addressing these challenges to enhance predictive capabilities, optimize network resources, and ensure robust, context-aware network operations requires using multimodal sensor data, such as vision, LiDAR, and wireless sensing, in novel artificial intelligence (AI)-based approaches. In this context, this introduces advanced multimodal collaborative perception frameworks utilizing state-of-the-art deep learning techniques, including Vision Transformers (ViT), convolutional neural networks (CNNs), and deep reinforcement learning (DRL), integrated within a digital twin framework. Specifically, the proposed MVX co-simulation environment serves as a digital twin that provides high-fidelity simulations mirroring real-world vehicular network conditions. This digital twin enables accurate predictions of line-of-sight (LoS) blockages, optimization of beam directions, dynamic channel prediction, and proactive resource allocation decisions. The results demonstrate the utility and practicality of digital twins for real-time network management and planning. Through extensive evaluations, this thesis demonstrates significant performance improvements over traditional wireless-centric approaches, achieving high prediction accuracy, reduced latency, increased spectral efficiency, enhanced energy efficiency, and robust URLLC performance. By combining multimodal sensing, advanced AI methodologies, and digital twin technologies, this thesis addresses key technical hurdles in next-generation vehicular networks and establishes foundational principles for intelligent, resilient, and context-aware 6G wireless communication ecosystems supporting critical applications such as autonomous driving and smart city infrastructure.

Date	7 Apr 2026
Original language	American English
Awarding Institution	École de technologie supérieure
Supervisor	Georges Kaddoum (Supervisor)