Researchers at the Massachusetts Institute of Technology have developed a system designed to help self-driving cars avoid accidental collisions when cornering, foreign media outlet Techspot reported. It is still a long way from the end product, but the progress they have made in recent years is promising.
The system, funded by the Toyota Research Institute, effectively analyzes changes in the shadows on the ground to determine if there are moving objects around the corner. It is based on an earlier version of the system, which was released at meetings in 2017 and 2018.
When testing with self-driving cars in the garage, the team’s modified system sensed approaching vehicles faster when sensing and parking, more than half a second faster than conventional lidar systems. Half a second doesn’t seem to make much difference, but in dealing with fast-moving self-driving cars, this may be the difference between a serious accident and a minor panic.
Other practical applications of the system include helping robots navigate busy corridors and even helping automatic wheelchairs navigate through hospital corridors.
But the researchers point out that the use of the system is currently limited. The researchers tested the system only in indoor environments with more consistent lighting conditions, which may make it easier for the system to detect and analyze shadows.
MIT has extensive experience in this field. The private research university has conducted experiments within its range, involving radar technology, artificial intelligence and even Wi-Fi to help it “see through” physical barriers.