Oxford smart cars take on US technology giants - Financial Times

©PA

Oxbotica’s technology is being used to power fully automated taxi pods for use in pedestrianised areas in the UK

Swerving around pedestrians and avoiding traffic cones, the sight of a two-seater self-driving pod navigating its way around the roads is not uncommon in the technological heartlands of California.

Except we are not in the US, but 5,000 miles away at a warehouse on the outskirts of Oxford where Oxbotica, a spinout from the nearby university, is showing off its latest gadget — a fully self-driving mini-car.

More

On this topic

IN Automobiles

The technology on display is not the car itself, but the robotic brain that runs it, called Selenium.

Oxbotica builds more than just the sensors and lasers needed to guide the vehicle along public roads — it makes the software that powers the entire system, processes the data and learns from mistakes to improve all the time.

This lends itself to uses other than in cars, with the group’s technology already in warehouse forklifts, mining lorries and even on Nasa’s Mars rover.

This breadth of uses will help it compete with larger and more established rivals in the market, such as Google, according to founder Paul Newman.

Oxbotica does not disclose accounts, but says it has made a profit every year so far. Although it has been an independent company for less than two years, Professor Newman says the group’s lofty ambition is to have its technology used in “all things that move”.

“We love those guys [Google], because they made the market possible,” he says. “They introduced this brilliantly into the public’s conception of what’s coming in the future.

“But first to the market does not equal first for all time. We’re talking about all things that move for all time. There’s not going to be one guy that does that for all time.”

He adds: “People say to us, ‘why would you do this if Google is out there? That’s like saying to Dell, why would you make computers when there’s IBM?’.”

Professor Newman sees cars, taxis, driverless pods and forklifts as “applications” of the underlying technology that can learn from the experiences of each other.

“If you had a prang, you are probably a different driver because of it,” says Professor Newman.

“But having an accident is a solitary learning experience. That’s not true when it comes to the machines. For every ding and every mistake, there’s the possibility that they all overnight learn from that. The ability for shared learning from those mistakes is something that machines offer that humans don’t have.”

The company is set to roll out driverless pods in Greenwich and Milton Keynes that use its systems under programmes part-funded by the UK government to spur homegrown development in driverless technology.

The investment drive, which will see up to £100m allocated to research projects, has also led to similar schemes in Coventry and Bristol.

Oxbotica has received several million pounds of funding. In Greenwich, London, the company’s technology will power driverless taxi pods for use in pedestrianised areas.

Because they learn from driving experiences, the cars will become increasingly smart in their environments as they become more familiar with the roads and conditions.

Oxbotica will still need to prove itself against other technology groups and carmakers striving to develop systems that allow vehicles to drive themselves in all environments and on all roads.

Google, in particular, has been testing driverless cars across California and has built up millions of driven miles of data. Other groups are busy mapping roads in detail to give autonomous systems comprehensive information on the world that surrounds them.

Ford and BMW are both working on driverless technology, which they say will be ready for commercial production next decade, while Volvo is working with Uber to develop self-driving taxis.

Uber last week rolled out its first fully driverless riding scheme in Pittsburgh, where customers hail a self-driving vehicle from their app. The company wants to log data from the trips to improve the learning ability of the cars.

But the road to driverless technology has not been without hitch.

Worries over the limitations of driverless technology have been raised by the disclosure of a recent fatal crash. A Tesla driver using the car’s autopilot system was killed when his vehicle collided with a van that its sensors failed to detect in bright sunlight.

Professor Newman calls the event a “tragedy”, but says “all sensible people” would expect there to be an accident sooner or later.

“We have never as a species built a tool that never had room for improvement on it.”

Copyright The Financial Times Limited 2016. You may share using our article tools.
Please don't cut articles from FT.com and redistribute by email or post to the web.

http://ift.tt/2cA1Dl2