Case Western Reserve University researchers share insight behind new SAE Edge Research Report on Off-Road Applications for AVs
Posted: September 30, 2024
Professor Kalle Lyytinen and Doctoral Candidate TC Eley are taking automated vehicles on the path less traveled—far less traveled.
Their SAE Edge Research Report "Implications of Off-road Automation for On-road Automated Driving Systems" explores how the complex and demanding nature of off-road environments yield benefits for not only off-road systems, but for the entire unmanned vehicle industry.
While this may seem like a leap for the technology, Lyytinen shared that the development of off-road autonomy is progressing faster than their on-road counterparts.
“Off-road vehicles are similar in some ways to on-road, but they observe the environment in a very different way. They need to explore it and be able to do very complex activities. Because of this, I think the innovation potential is much bigger than with on-road,” Lyytinen said.
With this greater potential for growth, vehicle developers need an accessible model to test application theory. Digital models, or “Digital Twins,” allow for simulations of off-road terrain to train autonomous systems to accurately detect even the smallest of nuances—increasing vehicle safety.
“If you have a smaller control environment, you to have a more precise understanding about the way that these systems are learning. Not only how they train, but how they learn while interacting with an environment. This wouldn't quite be possible within an on-road setting,” Eley said.
The authors identified a main challenge in the vehicle autonomy industry as “scattered information” regarding specialized operations per type of vehicle. They found that the findings of one field can lead to the growth of another, bridging these gaps of information. This can also support increased vehicle safety.
“With these off-road experimentations, we can figure out how learnings can be applied within different environments, then automate the aspects that may be too dangerous for a human while keeping aspects where they are most needed,” Eley said.
“Implications of Off-road Automation for On-road Automated Driving Systems” shows how off-road autonomy provides a rich environment for vehicle autonomy to accelerate—and provides researchers with a series of calls-to-action to help guide future research.
“The learning process is this: You try something, you learn, you get experience, you learn from it, and then you share it. This helps move the field ahead,” Lyytinen said.
Access to the report is available through an SAE Mobilus subscription.