Leveraging 5G and AI to Enhance Mcity Testing Capacity

Status: In Progress

Lead Researcher(s)

Henry Liu
Professor of Civil and Environmental Engineering, College of Engineering and Research Professor, Engineering Systems, University of Michigan Transportation Research Institute-Administration

Tinghan Wang

Huei Peng
Roger L McCarthy Professor of Mechanical Engineering, Professor of Mechanical Engineering, College of Engineering and Director of Mcity

Project Team

Tinghan Wang

Project Abstract

This project aims to combines 5G and AI technologies to further enhance Mcity testing capacity. Three topics are proposed.
1) Sensor-level augmented reality system with 5G and AI rendering technique. We will combine 5G, cutting-edging AI algorithms, and advanced simulation technique to generate sensor data for CAV tests. Two modes are proposed: purely-virtual AR and semi-virtual AR. The former simulates sensor data in various scenarios on a remote server and transfers the data to CAV by 5G in real-time. The latter uses real sensor data (i.e., camera and LiDAR) as the basis, and use AI algorithms to render the real data with various weather conditions and virtual objects, e.g., cars, pedestrians, bicycles, obstacles, and adversarial attacks. 5G communication is used for data exchange between CAV and a remote server that provides edge computing. The developed technologies will enable tests in various weather and scenarios, e.g., “testing winter in summer”.
2) Deployment of V2I2V in heterogeneous communication (5G and DSRC) at Mcity. We aim to leverage the advantages of both 5G and DSRC to design a fully connected Mcity. For example, the cars (BSM) and traffic lights (SPaT message) can be connected by DSRC as Mcity has done; while bicycle, pedestrian, infrastructure-based sensors (camera and LiDAR), and other testing facilities (e.g., streetlights, crosswalks, rail crossing gate) will be connected by the 5G network. A bridge between the cellular network and DSRC will also be developed to improve testing flexibility.
3) Develop self-driving systems to demonstrate the proposed AR and V2I2V systems. Live self-driving demonstrations using the MKZ platform will be developed to show how the sensor-level AR system, the 5G/DSRC mixed V2I2V, AI algorithms, and edge computing benefit CAV tests and safety in Mcity.

Project Outcome

This project aims to leverage 5G communication and AI algorithms to develop a sensor-level virtual testing environment and 5G/DSRC based V2I2V technology, in order to enhance Mcity testing capacity and improve self-driving safety.


BUDGET YEAR: 2020-01-01
IMPACT: SAFETY