ALLPCB
Summary
A white paper released by the Mcity testing facility at the University of Michigan states that augmented reality can increase the testing speed of connected and autonomous vehicles (CAVs) by 1,000 to 100,000 times and reduce the marginal cost of additional vehicle testing to near zero.
Testing Challenges for Autonomous Vehicles
Compared with conventional driver-operated cars and trucks, autonomous vehicle development poses additional challenges. Beyond verifying reliability and passenger safety, driverless vehicles must be able to prevent and avoid collisions, which requires testing large numbers of crash and near-crash scenarios, including events that are rarely encountered by conventional vehicles.
Augmented Reality Test Method
A research team at the U-M Transportation Research Institute (UMTRI) developed a novel test method that uses augmented reality to combine the real world with a virtual world, creating a faster, more effective, and more cost-efficient approach for testing CAVs.
Led by Professor Henry Liu, the team adapted video game and other virtual technologies to create an augmented reality environment. Real test vehicles inside U-M Mcity interact in real time with computer-generated vehicles via connected vehicle communication technology. Researchers can create test scenarios in the UMTRI Michigan Transportation Laboratory where physical test vehicles and computer-generated vehicles interact; the lab also serves as Mcity’s test control center.
How Virtual and Real Vehicles Interact
According to the white paper, observers can see a test vehicle approach a traffic signal and stop a few yards from the intersection to avoid rear-ending a computer-generated “virtual” vehicle stopped at the light. The computer-generated virtual traffic elements are transmitted using secure wireless technology to the Mcity test vehicle, allowing real and virtual vehicles to communicate with each other and with the test track infrastructure.
Researcher Comments
Professor Liu said, “To achieve widespread public adoption of driverless vehicles, it is necessary to first demonstrate that they are safe and reliable, which requires rigorous and extensive testing; otherwise deployment could take many more years. Virtual reality testing is not only more efficient and safer, it can also ensure autonomous vehicles operate reliably with the ability to prevent and avoid collisions. Our program has demonstrated that it can accelerate testing and reduce costs, and that a computer-generated library of traffic scenarios allows testing without damage or injury.”
