Drive.ai, an autonomous vehicle startup founded in 2015 by graduates from Stanford University’s AI Lab, has announced the launch of an on-demand driverless car service in Frisco, Texas.
The service will hit the road this July, and aims to be Texas’ answer to the ‘last mile’ transit problem: getting from A to B when it’s too far to walk, but too close to justify driving – and hunting for a parking spot. Using the Drive.ai app, locals will be able to hail one of the company’s self-driving vehicles on demand.
Starting with safety drivers
The company is launching a pilot scheme on public roads in the city between the offices of the Hall Group, and The Star, a venue with shopping, dining, and entertainment options. As with similar self-driving trials from the likes of Waymo and Uber, Drive.ai’s pilot will feature safety drivers at the wheel of every vehicle.
In a written statement, the company has outlined its next steps, should the pilot scheme be successful. “We plan to soon remove the safety driver in favour of a chaperone in the passenger seat, who will be educating riders about the self-driving experience and will attend to the vehicle. And eventually, we will offer passenger-only rides.
“Throughout every phase of the programme, our self-driving vehicles will be connected to our tele-choice technology, providing the ability to call a remote operator should extra assistance be needed.”
Choosing safety over subtlety
The Drive.ai team is aware that, in their words, driverless vehicles have “unique strengths and limitations”. Alongside the deep learning built into the entirety of the self-driving system – from perception and mapping to motion-planning and the mobile app – Drive.ai has decided to make its vehicles as conspicuous as possible.
All Drive.ai cars are painted bright orange. And though much more is needed to ensure safety than a garish coat of paint, the company believes that a distinctive colour scheme will help the public recognise driverless cars and respond to them accordingly.
Just as important are the exterior communication panels, which aim to convey each vehicle’s intentions to pedestrians and other drivers.
For the pilot scheme, the Drive.ai vehicles will be limited to a geofenced area. This will help the local team manage the fleet and move quickly to address any issues.
“Today’s announcement is bigger than just Frisco,” said Drive.ai, “it’s about moving the transportation industry forward, and deploying a Level-4 self-driving system to improve the state of mobility today.”
Internet of Business says
Any driverless fleet is only as good as the software and hardware built into each vehicle, and the Stanford AI pedigree bodes well for this comparatively late entrant, which aims to reach Level-4 autonomy more swiftly than some of its competitors. The conspicuous design is a sensible choice in these early days for autonomous vehicles, as is the information screen on the side of the vehicle.
But a question that is rarely, if ever, asked, is this: can AI and self-driving systems trained on America’s long, straight freeways and, in some cases, grid-like cities, be easily transferred to Europe or the UK, for example, where road systems are often more complex, challenging, and unpredictable?
While crude, the longstanding Top Gear joke about many American cars being unable to go around corners had a grain of truth to it in real-world tests. Conceivably, self-driving transport may face a long and winding road yet, outside of the big skies of Texas, California, or Arizona where many vehicles are being tested.