Google reveals self-driving car
(AFP) The tech titan that once upon a time was little more than a search engine has announced ambitious plans to build its own autonomous cars and it's so confident in the design and the technology underpinning them that the first models won't be offered with a steering wheel, or an accelerator or brake pedal.
In fact, all that the dashboard will contain is a screen showing the destination and a stop/start button. Instead, the car will use a combination of lasers, cameras, sensors and GPS to safely navigate the road ahead and to avoid other road users and potential hazards.
"Just imagine: You can take a trip downtown at lunchtime without a 20-minute buffer to find parking. Seniors can keep their freedom even if they can't keep their car keys. And drunk and distracted driving? History," said Chris Urmson, Google's Self-Driving Car Project Director, in a blog post announcing the new initiative.
Google has been developing autonomous driving technology for a number of years and its highly modified Toyota Priuses, Honda sedans and Lexus SUVs have become a common sight on the freeways in and around Mountain View, California, generating a huge amount of publicity in the process.
It is estimated that converting each of the cars in its self-driving fleet costs in excess of $150,000 on top of the list price of the initial vehicle.
And after 700,000 miles of testing, Google feels it's ready to take things to the next level. And that means building its own cars -- a fleet of 100 prototypes to be precise.
They will be powered by an electric motor with a limited range and a top speed capped at 25mph for safety. Taking tips from the Japanese school of vehicle design, the cars appear to be smiling, and the friendly face is intentional -- Google's way of making something technologically complex seem approachable and non-threatening. However, by stripping the controls back to a single stop/start button, it could be taking things too far.
There is little doubt that the vast majority of traffic accidents are caused by human error, or that removing humans from many driving equations would result in safer roads. But if the driver is to be eliminated, it has to be by a technology that is absolutely robust, that isn't going to crash and need rebooting and that is capable of processing all of the potential hazards that a car can encounter.
This is why so much research and development into autonomous driving technology undertaken by traditional carmakers -- from BMW and Mercedes to Ford and Volvo -- has been focused on maintaining the driver's attention and ensuring that he or she is still sufficiently alert to step in and take control when the car is unable to process the situation.
This unpredictability in terms of programming and in terms of driving situations is also why until this April, Google's own testing had been confined to the freeway where risks and issues can be quantified -- pedestrians and cyclists don't use freeways for instance and buses don't stop on it to drop off or collect passengers.
This is a point that Urmson conceded in April when the company announced its plans to start carrying out tests in town: "We still have lots of problems to solve, including teaching the car to drive more streets in Mountain View before we tackle another town, but thousands of situations on city streets that would have stumped us two years ago can now be navigated autonomously," he said.
However, he also revealed that the software that manages the car's controls was now capable of detecting "hundreds of distinct objects simultaneously -- pedestrians, buses, a stop sign held up by a crossing guard, or a cyclist making gestures that indicate a possible turn."
There is a huge interest for autonomous driving features on vehicles in the US where thanks to urban sprawl and city planning influenced by the big car makers, it is often impossible to walk even short distances.
The latest JD Power survey into Automotive Emerging Trends, published this month, highlights that 24% of US drivers would be willing to pay a $3,000+ premium for some form of autonomous driving feature on their next car (up from 21% in 2013's survey), whether that be self-parking or self-driving on freeways and in traffic jam situations.
However, drivers shouldn't get their hopes up too much yet. The prototypes are going to be tested by Google's own drivers and the cars are meant to serve as test beds, gathering data that will feed into the development of better and better iterations until the car is ready for the public, as Urmson explains: "If all goes well, we'd like to run a small pilot program here in California in the next couple of years. We're going to learn a lot from this experience, and if the technology develops as we hope, we'll work with partners to bring this technology into the world safely."
Ford, Volvo and Nissan are all also currently running large-scale tests of autonomous vehicles in real-world situations. In December, Volvo announced an initiative in partnership with the Swedish government that will eventually see 100 self-driving cars hit the streets of Gothenburg.
The large-scale project has been devised to study everything from infrastructure and sharing information between vehicles to the social, economic and environmental benefits or issues facing the technology, and the goal is to bring consumer-tested autonomous driving functions to its production cars by 2020.
Nissan has also pledged to offer self-driving cars by the same date and has built a dedicated test circuit for developing its technologies.