ºÚÁϸ£ÀûÍø

AUDRI conducts AI research in:

• Advanced, end-to-end, data-driven AI
• AI world models and reasoning models
• Perception, mapping, and localization
• Decision making, planning, and control
• Human-AI interaction
• AI testing, validation, and certification

AUDRI offers a unique, fully-integrated AI-research environment with:

• Full-stack, open-source E2E AD vehicles
• Driving data, E2E DNN models, and HPC
• Campus testbed with HD maps
• AI Digital Twin of the entire ºÚÁϸ£ÀûÍø campus
• Driving- and scaled-vehicle simulators

In this video, you can get a glimpse of the AUDRI AI lab and see what is happening in this episode of ºÚÁϸ£ÀûÍø's Stay ºÚÁϸ£ÀûÍøned.

AUDRI has AD-AI vehicles based on E2E Autoware, providing seamless ROS 2 integration in an L4 stack.

Autoware supports road-legal deployment, and ºÚÁϸ£ÀûÍø is an Autoware Center of Excellence. 

360â—¦ sensors

• Vision
• Radar
• Lidar
• Thermal

Naturalistic driving data – AI models

AUDRI collects real-world driving data on the ºÚÁϸ£ÀûÍø campus to support research on AD-AI architectures.

ºÚÁϸ£ÀûÍø has SPIKE1 - an AI supercomputer - for training AD-AI foundation models, and fine-tuned models are readily available.

Campus with HD maps  â€“  AI digital twin

ºÚÁϸ£ÀûÍø’s campus serves as a testbed for AD-AI vehicles, featuring HD maps and multi-modal localization.

A real-time AD-AI Digital Twin enables safe and efficient X-in-the-Loop (XiL) testing of edge cases.

Working with ºÚÁϸ£ÀûÍø in European projects is instrumental in achieving our Automated Driving goals.

Guus Arts
Project Manager, Advanced Technology, DAF Trucks N.V.