The researchers’ self-driving neural network gets into very few accidents in their virtual town.
Self-driving cars are powered by machine learning algorithms that require vast amounts of driving data in order to function safely. But if self-driving cars could learn to drive in the same way that babies learn to walk – by watching others around them and trying to mimic certain movements – they would require far less data. Hariri Institute Junior Faculty Fellow Eshed Ohn-Bar, an assistant professor of electrical and computer engineering, developed an efficient, safe, and collaborative paradigm for watching and predicting other cars’ actions to train autonomous vehicles.
Ohn-Bar and Jimuyang Zhang, a PhD student in electrical and computer engineering, presented their findings recently at the 2021 Conference on Computer Vision and Pattern Recognition (CVPR 2021).
The idea for the team’s training paradigm came from a desire to increase data-sharing and cooperation among researchers. Autonomous vehicles require many hours of driving data to learn how to drive safely, and some of the world’s largest car companies keep their vast data private to prevent competition. “There are a lot of autonomous driving companies and each company goes through the same process of taking cars, putting sensors on them, paying drivers to drive the vehicles, collecting data, and teaching the cars to drive,” said Ohn-Bar. Sharing driving data could help companies create safe autonomous vehicles faster, allowing everyone in society to benefit from the cooperation. Ohn-Bar highlights that no one company can solve this problem today on their own because current AI systems require so much data to work well. “Billions of miles are just a drop in an ocean of real-world events and diversity. Yet, a missing data sample could lead to unsafe behavior and a potential crash, and billions of miles are just a drop in an ocean of real-world events and diversity,” Ohn-Bar said.
The researchers’ proposed machine learning algorithm leverages data from other cars. The algorithm estimates the viewpoints and blind spots of other cars in the area to create a map that has a bird’s eye view of the surrounding environment. These maps help self-driving cars detect obstacles, like other cars or pedestrians, and can be used to understand how the other cars turn, negotiate, and yield without crashing into anything. A self-driving neural network is then trained by translating the actions of surrounding vehicles into the autonomous vehicle’s own frame of reference. These other cars may be human-driven vehicles without any sensors, or another company’s auto-piloted vehicles. Since observations from all of the surrounding cars in a scene are central to the algorithm’s training, this “learning by watching” paradigm encourages data sharing, and consequently safer autonomous vehicles.
Ohn-Bar and Zhang tested their self-driving cars in two virtual towns — one with straightforward turns and obstacles similar to their training environment, and another with unexpected twists like five-way intersections. In both scenarios, the researchers found that their self-driving neural network gets into very few accidents. With just one hour of driving data to train the machine learning algorithm, the autonomous vehicles arrived safely at their destinations 92 percent of the time. “While previous best methods required hours, we were surprised that our method could learn to drive safely with just ten minutes of driving data,” said Ohn-Bar.
While these results are promising, Ohn-Bar said that there are still several open challenges in dealing with intricate urban settings. “Accounting for drastically varying perspectives across the watched vehicles, noise and occlusion in sensor measurements, and various drivers is very difficult,” said Ohn-Bar.
The team’s paradigm for teaching autonomous vehicles how to self-drive could be used in other technologies as well. “This machine learning paradigm can be applied to many other systems,” said Ohn-Bar, “Delivery robots or even drones could all learn by watching other AI systems in their environment.”
Original Article: Cars that learn how to drive themselves by watching other cars
More from: Boston University
The Latest Updates from Bing News & Google News
Go deeper with Bing News on:
Training autonomous vehicles
- Cruise film stars self-driving car to build truston August 22, 2021 at 4:15 am
The car, one of 200 used for testing by the GM unit, stars in a short film meant to earn consumers' trust and educate them about the technology.
- Perrone Robotics Wins Autonomous Vehicle Shuttle Contracton August 20, 2021 at 9:25 am
Perrone Robotics Inc. has been awarded a contract from the University of Wisconsin-Madison to deliver an electric low speed vehicle (LSV) shuttle ...
- Tesla Bot: A concept humanoid robot embedded with Tesla's vehicle AIon August 19, 2021 at 7:15 pm
At Tesla AI Day, the company's CEO Elon Musk unveiled a concept humanoid robot that will be powered by the AI currently embedded within Tesla's autonomous vehicles. Aptly named the Tesla Bot, details ...
- University of Wisconsin-Madison and City of Racine Awards Perrone Robotics, Inc. New Autonomous Vehicle Shuttle Contracton August 19, 2021 at 3:02 pm
Perrone Robotics, Inc., a leading provider of fully autonomous vehicle (AV) technology and turnkey vehicle solutions for the mobility of people and things, was recently awarded a contract by the ...
- Apple won three 'Project Titan' patents today covering advanced LiDAR Systems for Autonomous Vehicles using Machine Learning+on August 17, 2021 at 6:02 am
Today, the U.S. Patent and Trademark Office officially published a series of Project Titan patents or Apple Inc. Two of the key patents cover LiDAR (also Lidar) systems. The first patent covers ...
Go deeper with Google Headlines on:
Training autonomous vehicles
Go deeper with Bing News on:
- Musk admits Tesla's latest self-driving software update 'not great'on August 24, 2021 at 7:45 am
Tesla is facing heat from regulators who question its Full Self-Driving feature's ability to live up to its name as Autopilot remains under investigation.
- Elon Musk admits Tesla’s new self-driving software ‘not great’on August 24, 2021 at 6:42 am
Tesla CEO Elon Musk admitted that the company’s driver-assistance software is “actually not great,” a week after federal regulators launched an investigation into Tesla’s Autopilot ...
- Tesla’s AI Day Made Self- Driving Technology Harder to Grasp. What Investors Need to Know.on August 24, 2021 at 5:30 am
Companies are racing to build autonomous vehicles, but judging who has the best systems has become nearly impossible for ordinary people.
- Elon Musk says Tesla's latest full self-driving software is 'not great' and that engineers are racing to improve iton August 24, 2021 at 5:02 am
Elon Musk said Tesla's partially automated full self driving (FSD) was "not great" - but that he'd tested the next version, and it was better.
- Elon Musk Just Tweeted a Brutal Admission About Tesla's Self-Driving Cars. Every Leader Should Be This Honeston August 24, 2021 at 1:04 am
If you want to know what Elon Musk thinks about something, there's a good chance there's a tweet for that. Usually, those tweets are good for their entertainment value, even if they don't always seem ...