Since introducing its autonomous vehicles, Google self-driving cars have traveled more than 1.7 million miles in a very public series of test drives and have been in only 11 accidents. Most of these collisions were a result of the driverless vehicle being rear ended by a human controlled car.
The human element is an important consideration in the development of autonomous vehicle technology. DSubhadeep Chakraborty, assistant professor of mechanical, aerospace and biomedical engineering, has focused his Joint Directed Research Development (JDRD) project on this element and is addressing it with connective vehicle technology, which asks the question “how can you drive as a group so that it is best for everyone?”
Connective vehicle technology, as the name suggests, allows the vehicles on a given section of road to communicate with each other, traffic signals and a variety of other sensors. This communication would allow vehicles to make real time adjustments to create the best possible driving environment with the other vehicles in range.
“Can we wake up 50 years from now and we are in a city with intersections without any street lights at all?” asked Chakraborty. “So what happens is as soon as I enter a zone surrounding the signal, my car starts talking to all the other cars that are approaching the intersection at the same time. And together through decentralized algorithms we figure out what is the best trajectory I can adopt and he can adopt so that together we can just go through the intersection without ever stopping.”
This lofty goal comes with a multitude of complex research question, not the least of which is how much distance should be maintained between the cars in order to keep the people in them from panicking. To answer these questions Chakraborty’s team is approaching this project with a multi-platform method.
The team has outfitted a donated Humvee with a series of sensors that feed data from the vehicle, as it is being driven, into a simulation they’ve created. The simulation allows Chakraborty’s team to study the data and use it to address the effect of the human element in connective vehicle technology.
“At the end of the day, transportation engineers want something that you can deploy, something that will become a reality, and this will not become a reality until you consider the human factors,” said Chakraborty. “So we need to bring that in and see how the controls can cope with that, because the human is not going to cope with the controls, the controls have to cope with the human.”