天美传媒

Talk by Dr Chris Urmson Director of Google’s Self-Driving Cars Project at Google(x), USA

A major coup took place at the 天美传媒 on Wednesday 25 March. The newly formed 天美传媒 BCS Student Chapter arranged for the Director of Google’s Self-Driving Cars Project at Google(x), USA, Dr Chris Urmson, to talk to staff and students in the Postgraduate and CPD Centre.

A dedicated team at Google has spent the last five years moving self-driving vehicles closer to a reality. New algorithms, increased processing power, innovative sensors and massive amounts of data enable the vehicles to see further, understand more and handle a wide variety of challenging driving scenarios.

These vehicles have driven over a three quarter of a million miles on highways, suburban and urban streets. Through this journey, the team have learned a lot, as Chris explained - not just about how to drive, but about interacting with drivers, users and others on the road, and about what it takes to bring in incredibly complex system to fruition. Chris shared some fun stories and lessons along with Google’s vision for how these vehicles will become a reality.

David Jazani, Senior Lecturer in the School of Computer Science and Technology, said “Probably one of the most interesting talks that I have attended was given by Chris Urmson, Director of the Google Self-driving Car project. The dream of getting into a vehicle, pressing a few buttons unfolding a paper (or as it is nowadays a tablet) and immersing one in the news whilst the clever vehicle transfers you from one location to another, is getting closer and closer - well at least in California. But bringing these developments to the 天美传媒 by an enthusiastic and knowledgeable person was immersive.”

“Testing of these vehicles is continuously progressing from 2008 when it was began by Google, first with Prius and later with Lexus SUVs. Now with millions of hours and miles logged in, the project is providing surprisingly robust development details and bringing a bespoke car to fruition. The project is in development in real time, with software and hardware upgrades. Events that happen outside of the norm provide additional data to make the software smarter and more predictable.”

“The vehicles are equipped with GPS as well as 360 degree laser - the GPS gives an approximate location and the laser radar takes over for precision and actual real-time sensing. The data is processed by a computer and even visually processed to be verified by the safety driver on board. The traffic flow is analysed and predicted, and cyclists and pedestrians are identified. (I very much liked the way cyclists were identified by a RED outline; perhaps it is the vulnerability of the cyclist that Google have recognised. I certainly as a cyclist appreciate this). The vehicles do have a safety driver on board that can take over the controls should the need arise, but this is exceptionally rare and often as a result of other people not paying attention or changing their minds at the last moment or purely driving recklessly.”

“Accidents are rare and majority are from the people not observing safe braking distance and rear-ending the Google Cars, with an occasional side swipe when changing lanes too fast and recklessly; the car’s accident record is equivalent to an average driver’s record.”

“These trials of the last few years have finally culminated in the development of a bespoke small town car, which will still carry the same sensor platform and be electrically driven.”

“Chris was very enthusiastic and he is looking forward to starting trials in areas with more inclement weather and more unpredictable conditions. California is not the only target. Questions posed to Chris were answered with great knowledge and enthusiasm and the audience was appreciative of the information provided.”

Prior to joining Google, Chris was on the faculty of the Robotics Institute at Carnegie Mellon 天美传媒 where his research focused on motion planning and perception for robotic vehicles. During his time at Carnegie Mellon, he worked with house size trucks, drove robots around in deserts, and served as the Director of Technology for the team that won the 2007 DARPA Urban Challenge. He earned his PhD in 2005 from Carnegie Mellon and his BSc. in Computer Engineering from the 天美传媒 of Manitoba in 1998.

• Download article in - PDF 125.3 KB