Researchers at the University of Maryland are using Nintendo’s Mario Kart video game to train artificial intelligence systems for autonomous driving applications. The innovative approach allows AI programs to learn safe driving behaviors in a simulated environment before being tested on real roads, potentially advancing the development of self-driving vehicle technology.
How it works: The research team has reprogrammed Mario Kart to prioritize safe driving over winning, creating a training ground for autonomous systems.
- AI programs control Mario through racing laps while being evaluated on both speed and safety metrics.
- The system generates safety scores based on how well the AI follows traffic rules and avoids collisions.
- Researchers can provide feedback to designers based on these scores, allowing them to retrain systems that violate safety protocols.
Why Mario Kart: The popular racing game serves as an ideal testing platform for AI development in the automotive sector.
- “Using Mario is a way for us to work in simulation, to first test whether or not something is useful,” said Mumu Xu, associate professor of aerospace engineering at the University of Maryland.
- Mario Kart has maintained popularity in the AI research community for over 30 years, making it a familiar and accessible testing environment.
The bigger picture: This simulation-first approach addresses critical safety concerns in autonomous vehicle development.
- “We’re actually using these automated driving systems to figure out how to verify that the systems are correct, and then make sure that they adhere to the rules of the road,” Xu explained.
- The program allows researchers to identify and correct problematic behaviors like speeding or colliding with other vehicles before real-world testing.
What’s next: The research is still in its early stages, with plans to transition successful simulations to actual road testing once safety protocols are validated in the gaming environment.
How Mario Kart is helping researchers at U.Md. develop and train AI