LEADING THE MOBILITY TRANSFORMATION

Mcity 2.0

Mcity > Mcity 2.0

The Mcity Test Facility has evolved into the next-generation autonomous vehicle test track thanks to a $5.1 million grant from the National Science Foundation. Mcity 2.0 gives researchers easy, remote access to CAV testing resources, and will help create a more equitable playing field in mobility. Learn more about Mcity 2.0 and submit your request to use it by clicking the buttons below.

SUMBIT REQUEST
USING MCITY 2.0
AV CHALLENGE

NSF National Science Foundation logo
About Mcity 2.0

With NSF’s support, Mcity Director Dr. Henry Liu,  Managing Director, Greg McGuire, and the Mcity engineering team enhanced the Mcity Test Facility by developing digital infrastructure to overlay the physical test facility and create a cloud-based, augmented-reality CAV testbed available to academic and industry researchers nationwide.

Mcity 2.0’s remote capabilities are now operational, and can be used with the physical test facility and Mcity research vehicles. 48 test days per year have been allocated for NSF projects. Mcity expects to accept about 10 new projects annually.

Mcity 2.0 Enhancing Autonomous Vehicle Testing

Mcity 2.0 is a platform designed to help users remotely test and refine autonomous vehicle (AV) motion planning algorithms. It allows users to conduct tests in both simulated and mixed reality environments without needing a complete AV system or physical testing facility. Mcity 2.0 integrates digital infrastructure with physical facilities to provide real-time visualization of AV status, safety metrics, and testing data, which are archived for ongoing analysis and improvement.

Goals and Objectives

  1. Create realistic traffic environments for AV testing.
  2. Enable remote access and control of AV testing facilities.
  3. Offer comprehensive evaluation metrics and sensor data to enhance AV motion planning algorithms.

Key Features

  • TeraSim on the Cloud: A fast, realistic traffic simulation generating critical events like collisions. It uses models calibrated with real-world data to simulate both normal and safety-critical driving scenarios.
  • AWS Cloud Hosting: TeraSim is hosted on AWS, allowing global remote access for seamless integration with user algorithms via the TeraSim API.
  • Evaluation Metrics: Includes minimum distance, time-to-collision, and speed difference to assess AV safety and performance.
  • Remote Control Communication: Uses Redis for data exchange and a low-latency communication pipeline to manage real-time data and control commands.

Testing Capabilities

  • Simulation-Based Testing: AVs and background traffic are visualized in real-time, with performance metrics plotted. Data is archived for post-analysis.
  • Mixed Reality Testing: Combines digital simulations with physical AVs. Remote testers can view real-time AV status and control inputs via Mcity’s HD map.

Testing Procedure

  1. Simulation Testing: Remote visualization and data recording of AV and virtual traffic.
  2. Mixed Reality Testing: Real-time visualization of physical AV’s control inputs and performance on Mcity’s map, with remote access to onboard and roadside views.

Outcomes

  • Simulation Testing: Identified potential safety issues and improved longitudinal control to reduce conflicts.
  • Mixed Reality Testing: Enhanced passenger comfort by smoothing deceleration and reducing harsh braking.

By leveraging Mcity 2.0’s advanced simulation and mixed reality capabilities, remote users can effectively test, validate, and enhance AV motion planning algorithms, driving progress in autonomous vehicle safety and performance.