By Jonny Lupsha, Wondrium Staff Writer
A fatal 2018 car crash involving driverless software has been resolved, BBC News reported. New evidence shows the driver was playing a smartphone video game, contrary to earlier concerns over the semi-autonomous software in the car. So where do robotics in cars currently stand?
According to the BBC News article, the U.S. National Transportation Safety Board (NTSB) has published the results of a two-year investigation into the death of Walter Huang, 38, whose Tesla merged into the “gore area” of a highway—the paved stretch between two forking branches of highway. The results stated that Huang failed to take control of the car because he was distracted by a smartphone game and the Tesla’s collision detection system wasn’t designed to account for the barrier into which he crashed.
Self-driving automobiles have been a part of the cultural zeitgeist at least since old science fiction stories depicted them in the 20th century, if not earlier. However, practical work on them has increased in recent years.
Stanley and the DARPA Grand Challenge
Until this century, nobody had really tackled the idea of getting a driverless car to travel from point A to point B.
“Combining navigation with autonomous driving in robotic cars was the primary challenge put forward in 2004 by the Defense Advanced Research Projects Agency of the U.S. Department of Defense, which is commonly known as DARPA,” said Dr. John Long, Professor of Biology and Professor of Cognitive Science on the John Guy Vassar Chair of Natural History at Vassar College.
According to Dr. Long, DARPA called it the DARPA Grand Challenge. The goal was for an unmanned, autonomous vehicle to navigate a 150-mile route with no external assistance. No vehicle completed the 2004 Grand Challenge, but the following year, DARPA re-ran the challenge. Five cars completed it, but the winner was a robot named Stanley, a modified VW Touareg that was built by a robotics team at Stanford University. The team used three sensor types to help Stanley navigate on its own.
“Radar was the longer-distance system,” Dr. Long said. “It’s good for keeping track of other vehicles and is insensitive to things like shadows that can throw off a video system.”
Dr. Long said that Stanley’s second system used LIDAR, named for “light” and “radar.” “The light of a LIDAR is laser light, which the sensor shines out into the world and then reads the radar-like rebound of that light to measure distance,” he said. Finally, a video camera used optic vision to provide further input.
After DARPA
According to Dr. Long, the 2005 DARPA Grand Challenge was seen as a major milestone in robotic driving, and the world noticed. Several of the Stanford robotics team members took some of their achievements with Stanley to Google, he said, and began work on Google Driverless.
“Nearly all automobile manufacturers have begun work on autonomous, driverless cars. They have been helped immensely by components manufacturers like Bosch, who are making robotics systems that can be deployed on any vehicle. Like Google, Bosch’s driverless system uses LIDAR, video, and radar to to create a dynamic map of the world.”
Much of the technology being used in driverless cars, Dr. Long said, can also be applied to trains, trucks, and buses. This offers the enormous potential of completely overhauling our transportation networks and drastically reducing vehicular deaths. The transition will take some time, but it’s far from a pipe dream.
“The technology is already available to make robotic cars not only feasible, but also safer and more efficient, an improvement that will probably be rewarded with reduced insurance costs,” Dr. Long said. “Robotic cars may overcome the trade-off of safety vs. speed.
“As the infrastructure for robotic cars to communicate with each other through intelligent networks continues to develop, it’s even possible that some roadways will be restricted to robotic cars only, offering passage that’s fast and safe.”
Dr. John Long contributed to this article. Dr. Long is a Professor of Biology and a Professor of Cognitive Science on the John Guy Vassar Chair of Natural History at Vassar College. He received his Ph.D. in Zoology from Duke University.