Vehicles · Drones · Intelligent transportation
Our V2X lab integrates AI, V2X communications and digital twins across system and network research to advance connected, automated mobility and intelligent transportation.
Vision-language models (VLMs), multimodal sensor fusion and multi-agent learning connect embodied perception with reasoning and action. Key directions include spatial scene understanding, object tracking and motion prediction across vehicles, drones and roadside infrastructure.
Cooperative perception, trajectory prediction and distributed planning underpin automated driving in shared road environments. Integrating ETSI cooperative awareness and collective perception messages with multi-agent decision-making enables research on manoeuvre coordination, intersection negotiation and cooperative merging under communication delays and partial observability.
Coupled mobility–network digital twins integrate traffic microsimulation, sensor models, vehicle dynamics and wireless communication. Closed-loop co-simulation supports scenario generation and reproducible evaluation of cooperative perception, manoeuvre planning and network performance under varying traffic, channel and latency conditions.
Cross-layer resource allocation, MAC scheduling and congestion control address connectivity challenges in ITS-G5 and 5G NR-V2X, complemented by RAN slicing in cellular networks. Emerging 6G research encompasses integrated sensing and communication (ISAC) and reconfigurable intelligent surfaces (RIS), with emphasis on beamforming, RIS configuration and sensing–communication trade-offs under high mobility.
V2X-assisted cooperation connects onboard intelligence with roadside edge processing and cloud learning. Key research directions include edge-assisted multi-vehicle sensor fusion, cooperative inference and mobility-aware task offloading. Joint communication–computation scheduling addresses information freshness and end-to-end latency constraints for collective perception and manoeuvre coordination.
Semantic communication selects and compresses task-relevant information for cooperative perception and manoeuvre coordination. Research targets lower bandwidth use and less redundant messaging while preserving information freshness, task accuracy and compatibility with ETSI V2X services.
Keeping safety messages timely and reliable when many vehicles share the same wireless channel.
Sharing vehicle positions and intended paths to coordinate safe crossings, even when messages arrive late.
Delivering hazard warnings to nearby vehicles quickly without flooding the network with repeated messages.
Sharing vehicle and roadside observations to track pedestrians and cyclists when their position or movement is uncertain.
Choosing when and where vehicles transmit to reduce interference and missed safety messages.
Recreating roads, vehicles and wireless links to test how delayed or lost messages affect cooperative driving.
Allocating wireless capacity as traffic changes so safety messages remain timely alongside other connected-vehicle services.
Combining onboard sensor data with information from nearby vehicles to understand complex road situations.
Sharing data processing across vehicles, roadside computers and the cloud to support timely driving decisions.
Sending only the information needed for cooperative driving, reducing message size without losing important details.
Using observations shared by nearby vehicles and roadside sensors to detect road users hidden from view.
Testing V2X under different traffic, driving and wireless conditions to identify when cooperative driving could fail.
Maintaining V2X services as vehicles move between network coverage areas and roadside computing systems.
Sharing observations and intended paths to support safe merging and collision avoidance despite communication delays.
Research towards standards
Our research studies contribute to the development of standards such as ETSI and 3GPP, through protocol design, reproducible evaluation and evidence.
