Dubbed Waabi World, the simulated driving school for autonomous vehicles is the first offering from the high-profile startup founded by Canadian AI star Raquel Urtasun. It aims to help engineers develop the brains of self-driving software while removing the on-road risk
A new solution to autonomous driving safety has been born in Toronto with an announcement today from artificial intelligence firm Waabi about the launch of its first project: Waabi World.
The concept is pretty simple: a driving school for autonomous vehicles that allows them to build up experience, data and “brain power” before they even hit the road.
The execution of Waabi World though, is anything but easy.
Drawing on the most advanced technologies and software expertise, Waabi — founded in 2021 by Uber’s former chief scientist for autonomous vehicle efforts, Raquel Urtasun — is attempting to create a digital environment that will cultivate a human-level intelligence in autonomous driving software by stress testing the vehicle’s software to identify areas of weakness and, eventually, train the machine to be able to think for itself.
Imagine Waabi World as a video game, says the company, that constantly plays against the Waabi Driver.
“When you get behind the wheel, a mix of intuition, instinct and learned skills helps you process what’s happening and make instantaneous decisions about how to navigate obstacles, when to slow down, speed up, stop, and much more. The human brain’s ability to do all this is remarkable,” reads the Waabi website.
“Waabi World exposes the Waabi Driver to the vast diversity of experiences needed to hone its driving skills, including both common driving scenarios and safety-critical edge cases. This significantly reduces the need to drive testing miles in the real world and results in a safer, more affordable solution.”
Waabi World is designed to be used in trucking applications as that industry is making an aggressive push towards achieving autonomous middle-mile and drayage capabilities and is routinely plagued by driver shortages and supply chain disruption.
Waabi the company is a recent arrival on the autonomous driving scene, but Urtasun has had a high-profile figure ever since she was recruited to Uber in 2017 from a professor position at the University of Toronto. When that disbanded in late 2020, she began work on Waabi, landing a massive US$83.5 million in early-stage funding from leading Silicon Valley and Canadian venture firms a few months later.
There’s little doubt this first unveiling will command attention.
Waabi World took just six months to develop, said Urtasun in an interview with BNN Bloomberg. One of the most valuable of elements of the program is that it evaporates the amount of time required to train the AV software to change its approach in response to on-road problems.
The core capabilities of Waabi World, per the company website, are:
- Builds digital twins of the world from data, automatically and at scale;
- Performs near real-time high fidelity sensor simulation enabling testing of the entire software stack in an immersive and reactive manner;
- Creates scenarios to stress-test the Waabi Driver, automatically and at scale;
- Teaches the Waabi Driver to learn from its mistakes and master the skills of driving without human intervention.
“Oftentimes you will see companies say, ‘We drive millions of miles in the real world’ and that’s actually in my opinion a negative, because they require all the driving in order to really understand or see when the system fails,” said Urtasun to BNN.
“It’s like finding the needle in the haystack … and you don’t want to wait to the real world to see a problem.”
Thanks to an immediate feedback loop, Waabi World is able to transmit data back to the autonomous software in real time regarding any corrections in behaviour needed. The software can then be put to the test again right away using the same conditions, which helps engineers to measure the software’s learning curve. Where in the real world it could take 100,000 miles to “teach” autonomous software how to respond correctly to the same problem, in Waabi World that could be shortened to mere hours.
Urtasun says she does not intend to make Waabi World available to the wider autonomous driving industry.
Waabi runs its own small fleet of autonomous trucks and Waabi World will be tested alongside that on-road fleet to compare how the virtual driving school does against the school of life.
Eventually Waabi World may be used by regulators to help assess and evaluate future autonomous driving technology in a safe and expedited way, but that use case is a long way off. For now the company is single-mindedly focuses on its top priority: making autonomous driving safer.
“It might seem counterintuitive, but we want to see the Waabi Driver fail. We don’t want to wait until we test in the real world to see the system failing,” reds the Waabi World website.
“Self-driving is one of the most exciting and important opportunities in technology today. Once realized and scaled, it will change life as we know it — how we operate businesses, power industries, build cities, and move goods and people.”