Mcity 2.0 is the next-generation autonomous vehicle test track. While capabilities are not yet fully operational, Mcity’s physical test facility and research vehicles are available now to researchers funded by the National Science Foundation. Learn more about Mcity 2.0 and submit your request or questions below.

NSF National Science Foundation logo
About Mcity 2.0

The Mcity Test Facility opened in 2015 as the world’s first purpose-built proving ground for testing the performance and safety of connected and automated vehicles (CAVs) and technologies under controlled and realistic conditions.

Today, the facility is evolving into the next-generation autonomous vehicle test track thanks to a $5.1 million grant from the National Science Foundation. With NSF’s support, we’re enhancing the Mcity Test Facility by developing digital infrastructure that will overlay the physical test facility and create a cloud-based, augmented-reality CAV testbed that will be available to academic researchers nationwide.

Mcity 2.0 will give researchers easy, remote access to CAV testing resources, and will help create a more equitable playing field in mobility.

Funding to use Mcity 2.0 will be available to researchers already working on NSF projects who seek additional support.

 

Mcity 2.0 Diagram which features the 4 pillars to the platform

Mobility Data Center

Real-time municipal traffic data feeds Mcity’s Mobility Data Center.

The Mobility Data Center is a cloud-based data platform designed to ingest, archive, process, and share the large amount of data generated by cooperative infrastructure located within the Mcity Test Facility and the City of Ann Arbor.

  • Hosted on Amazon Web Services (AWS)
  • Website user interface allows for real-time monitoring, near-miss/crash event querying, and raw data querying
  • Data is collected from roadside sensors (cameras) that capture images (vehicles and non-motorized road users, such as pedestrians and cyclists as well as motorcyclists and persons with disabilities or reduced mobility and orientation)
  • The images are then processed for tracking information, trajectory prediction, and timing of potential collisions using machine learning
  • Potential collision warnings are generated and sent to the roadside unit to be communicated to relevant vehicles
  • Near-miss and crash events are archived in the Mobility Data Center
  • Data can be used to train automated vehicles and has Smart City applications
  • Richer data and easier to collect than on-road driving data

Learn more about the Mobility Data Center

Naturalistic Driving Environment

In a 2016 report, the Rand Corp. estimated that it would take a fleet of 100 automated vehicles driving at 25 mph for 5 billion miles to demonstrate their ability to avoid crashes and fatalities in the real world. Computer simulation is vital for researchers to put their systems to the test, but access is limited in the academic realm. To date, eight commercial simulation vendors have built digital twins of the physical Mcity Test Facility that will be unlocked as part of the Mcity 2.0 rollout. This environment is also adversarial, meaning researchers can test against safety-critical scenarios such as hard braking, cut-ins, and more.

Source: How Many Miles of Driving Would It Take to Demonstrate Autonomous Vehicle Reliability? 

Learn more about the Naturalistic Driving Environment (NDE)

Mcity OS

Mcity OS makes it possible for researchers to create and execute complex, sophisticated, and easily repeatable test scenarios of connected vehicles, automated vehicles, and connected and automated vehicles – potentially saving testing time and costs, and accelerating product development.

Mcity OS runs on any internet-enabled device to control all the features of the Mcity Test Facility. This cloud-based open-source operating system gives users point-and-click control over interactions between vehicles and facility features and infrastructure.

Mcity OS tools can be integrated at other test facilities and in real-world environments as we lead the transformation to CAVs, CV2X, and smart cities. Mcity OS is also available for use at other test facilities. The American Center for Mobility is the first facility to license Mcity OS.

Mcity OS will be the tool researchers use to access Mcity 2.0.

Learn more about Mcity OS

Live Test

Mcity 2.0 will enable U.S.-based researchers  to send their test algorithms and programs to Mcity, plug them into the test facility’s Mcity OS operating system, request specific conditions for testing and participate remotely as those parameters play out in Mcity’s combined real/virtual setting

An example of the kind of AV testing that could be conducted using this approach is the Mcity Safety Assessment methodology outlined below, which was developed by Mcity and the Center for Connected and Automated Transportation (CCAT), a regional transportation research center funded by the U.S. Department of Transportation. CCAT is based at U-M.

Mcity Safety Assessment

  • The Mcity Safety Assessment combines two key pieces of technology in the Mcity 2.0 toolkit: the Mcity ABC Test and the CCAT SAFE TEST.

Mcity ABC Test

The Mcity ABC Test consists of three parts: Accelerated evaluation, Behavior competence, and Corner cases. Taken together, they are random, valid, fair, and comprehensive.

  • Accelerated Evaluation: In this process, the first step is to collect naturalistic driving data that reflects what the test vehicles will face on public roads in normal conditions. The behavior of the human drivers is then “skewed” (based on importance sampling) to boost aggressive/risky behaviors and focus on difficult miles. This process is possible because the University of Michigan and, in particular, the U-M Transportation Research Institute, has 20 years of experience leading field operational tests and has collected tens of millions of miles of naturalistic driving data.
  • Behavior competence: In behavior competence testing, vehicles are put through a set of comprehensive scenarios to demonstrate their safety performance. In collaboration with researchers from the University of Michigan and the Mcity Leadership Circle, 50 scenarios have been compiled.
  • Corner Cases: There is no official definition of corner cases, other than the fact that most experts agree these cases should be deterministic and should be at the very edges of the operating domain design. A corner case, for example, would involve an automated vehicle traveling at its highest speed with the most obstructed view, and facing a pedestrian running at the highest speed, and so on. Corner cases can also be designed to explore the known weaknesses of the subsystems of the vehicle, such as asking a vehicle relying on cameras to recognize a black car on a dark night.

CCAT SAFE TEST

The Safe AI Framework for Trustworthy Edge Scenario Tests, or SAFE TEST, combines two pieces of technology: the Augmented Reality Test Environment and the Naturalistic and Adversarial Driving Environment (NADE). It is estimated that one mile on a test track using the SAFE TEST equates to 5,000 real-world miles.

  • Augmented Reality Test Environment: To add real vehicles to a test scenario, companies would have to spend thousands of dollars and hours to coordinate and control. The augmented reality environment allows researchers to add virtual background traffic to Mcity that the test vehicle views as “real.” The simulated vehicles are easily controlled so specific test scenarios can be repeated perfectly each time.
  • Naturalistic and Adversarial Driving Environment: The NADE inserts background vehicles that conduct adversarial and rare maneuvers, such as hard braking, at a much higher rate. Simultaneously, the environment uses naturalistic driving data from the University of Michigan Transportation Research Institute (UMTRI) to ensure unbiasedness.

Learn more about live testing with Mcity 2.0

Mcity 2.0 Proposal Form

Read the Mcity Test Facility Overview

Read the Mcity Vehicle Features Document

Click or drag a file to this area to upload.
PDF format only. File size not to exceed 10 MB.