Why it matters: Researchers at MIT have developed a system aimed at helping autonomous systems avoid unintended collisions when rounding corners. It's far from a finished product but the advancements they've made in recent years are promising.
The system, funded by the Toyota Research Institute, effectively analyzes changes in ground shadows to determine if there is a moving object around a corner. It builds on an earlier version of the system that was presented at conferences in 2017 and 2018.
In testing with an autonomous car in a parking garage, the team’s revised system was faster at sensing and stopping for an approaching vehicle, besting a traditional Lidar system by more than half a second.
Fractions of a second may not seem like much to get excited about but when dealing with fast-moving autonomous vehicles, it could be the difference between a serious accident and a minor scare.
Other practical applications include robots navigating busy hallways and even autonomous wheelchairs traversing hospital corridors.
While it could one day be a helpful asset, right now, its use is limited. Researchers have only tested the system in indoor settings where lighting conditions are more consistent. This makes it easier for the system to detect and analyze shadows, we’re told.
MIT has plenty of experience in this field. The private research university has multiple experiments under its belt involving radar technologies, artificial intelligence and even Wi-Fi to help it see through physical obstacles.
https://www.techspot.com/news/82526-mit-helping-autonomous-vehicles-see-around-corners.html
