The Mcity Test Facility has reopened for contactless testing. Please e-mail Vaughn Haack for test facility information at The Mcity Office Building is still closed and Mcity staff will continue to work from home due to the ongoing coronavirus pandemic. Please feel free to contact us with any questions or concerns.

Mcity ABC Test: Develop and Implement 16 Behavior Competencies

Status: Complete

Lead Researcher(s)

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

Greg McGuire
MCity Lab Director and Adjunct Lecturer in Center for Entrepreneurship, College of Engineering

Project Team

Project Abstract

This project is a three-pronged approach for on-track testing of highly automated vehicles (HAVs); the three prongs are Accelerated evaluation (A), Behavior competence (B), and Corner cases (C). The first prong is for common driving scenarios such as car-following, cut-ins, and left turns, and the scenario parameters are obtained through importance sampling theory to make sure the risky/challenging cases are emphasized. This, instead of uniform sampling, can significantly reduce testing time. The primary focus of this project will be the behavior competence prong.
This project aims to realize 16 behavior competence scenarios completed in the Mcity test facility.
The key challenges involve designing the stochastic test cases and executing the test cases accurately and reliably using an Mcity CAV test vehicle and moveable platform that can carry a pedestrian, bicycle or robotic deer.

Project Outcome

Successfully demonstrate 16 behavior competence tests at the Mcity Test Facility, with the scoring system defined.
The Mcity ABC test is one of first testing concept for highly automated vehicles in the public domain. The project helps to form consensus and best practice.
Successful execution of the project will help to improve transparency and public trust toward highly automated vehicles.
A demo route was designed to showcase 5 of the 16 test scenarios.
A library of test cases (with specific risk levels: easy, moderate, hard) were generated.

BUDGET YEAR: 2019-01-01