To be held on November 3, it will provide a simulated ‘city streets’ environment, where Stanford’s latest robot car, ‘Junior’, will have to navigate streets filled with stationary and moving cars, obey road rules, traffic lights, road signs and road markings – but it will be in a controlled test environment, not on the streets of New York or London.

That’s why scientists say robot cars won’t be with us until 2030, because at today’s level of artificial intelligence, today’s robotic cars simply aren’t up to the task of driving themselves unaided on busy city streets.

Of course, a self-piloting car has been a dream for decades, but it’s only now that computing technology is finally starting to catch up – and there’s still a long way to go. Junior is set to use Intel Dual and Quad Core chips, along with several cameras, sensors, lasers and lots of software to try and understanding the hectic environment of real city streets – instead of just observing the environment and reacting to it as the Nevada race ably demonstrated.

After all, as Mike Montemerlo, a senior research engineer at the Stanford Artificial Intelligence Lab (SAIL) says, “This has a component of prediction. There are other intelligent robot drivers out in the world. They are all making decisions. Predicting what they are going to do in the future is a hard problem that is important to driving. Is it my turn at the intersection? Do I have time to get across the intersection before somebody hits me?”

What ties Junior’s artificial intelligence together is the software, which is being created by a dozen students, faculty and researchers at SAIL. Modules for tasks such as perception, mapping and planning give Junior the machine-learning ability to improve its driving and to convert raw sensor data into a cohesive understanding of its situation.

According to Montemerlo, new software development began last fall, with some testing of the team’s software modules in simulated traffic situations since the beginning of the year. The team expects to move into full-time testing and iterative improvement by the end of
March.

A self-piloting car, especially one that is powered by electricity, could become a mobile office, complete with all forms of onboard entertainment with Internet access and computing facilities, food storage and quick preparation facilities for snacks, cold drinks and coffee and even an inbuilt toilet – and would make rush hour traffic jams much easier to deal with.

But before all of this becomes reality sometime in the 2030s, we’ll all need to see solid proof that robot cars are faultlessly reliable, and that there is, at the very least, a set of easily accessible manual controls if required. After all, I don’t think I’m ready to rely on Windows Vista to ensure my safety on the road – or even Linux or Mac OS X for that matter – at today’s level of development and reliability, and I don’t think you are, either.

While computer crashes are tolerated today, having the same with a robot car certainly won’t be. Reliable robotic technology could reduce the road toll to virtually zero, and it’ll be a great day when it comes. But until then, much work remains to be done. Let’s hope we achieve it all before the asteroid gets here in 2036.
{moscomment}