Autonomous cars can currently perform most of the basic driving tasks, like switching lanes or parking, but avoiding an oncoming vehicle or avoiding an animal on the road is still too complex for most systems.
That is why Georgia Tech has built two rally trucks, 20 percent the size of normal trucks, to test some of the irregular events that happen on roads.
A team of researchers from the Daniel Guggenheim School of Aerospace Engineering and the School of Interactive Computing worked on the autonomous track, called the AutoRally.
Reaching a top speed of 20 mph – the equivalent of 90 mph at full size – the team could test multiple events at different speeds. The cars start to learn and adjust to the barriers on the road, learning how to avoid them.
The AutoRally trucks are fitted with a CPU, GPS, battery, and swarth of cameras and sensors to calculate “up to 2000 possibilities in 50 milliseconds.” The team’s goal is to make the truck react as quickly as a human would to complex situations.
Autonomous tech still not up to quick reflexes?
The inability to react to an unfamiliar situation is one of the main reasons people are still unsure about autonomous cars. In a new study from the University of Michigan, public perception of self-driving is still quite bad, with a lot of people cautious or against computers controlling cars — a study in the U.K. produced similar results.
Georgia Tech already provides all of its research and code online, allowing companies currently testing autonomous cars to try out the system in regular cars.
Google, the company that has worked the longest on self-driving, has been actively avoiding crashes, at least in public. It reported two crashes in California last month, both of which were caused by human error, though we have seen evidence of the autonomous car causing a crash.