Wayve expands AI ADAS testing to US

UK driverless car tech developer Wayve is starting on-road testing of its AI-based ADAS software in the US.
London-based Wayve has opened an office in Silicon Valley to support its US expansion and AI development with a testing program in San Francisco focusing on Advanced Driver Assistance Systems (ADAS).

The company is developing AI software capable of enabling a range of driving assistance and automation features for vehicles worldwide, with US tests complementing ongoing automated driving trials in the UK.

This marks Wayve’s first on-road trials outside the UK and an important step in developing AI software capable of powering various driving assistance and automation features for vehicles worldwide. The testing will help expand the geographic scope of Wayve’s AI Driver software with a focus on software development, hardware integration, and deployment operations.

UK looks to sovereign driverless car supply chain
Europ’e first commercial robocar hire service
Tesla’s robot taxis aren’t coming quite yet
“This marks a major milestone for Wayve. We are now testing our AI software in real-world environments across two continents. After years of pioneering a data-driven mapless AI Driver that can generalize across different vehicles and cities in the UK, we are excited to bring this technology to the US. San Francisco’s unique driving conditions offer rich data insights that will be crucial in further developing a global AI platform for automotive customers,” said Alex Kendall, Co-founder and CEO of Wayve.

The Foundation Model can provide the core intelligence needed to operate vehicles in open environments and support all levels of autonomy. The US tests cover Level 2+ (“Hands-off, eyes-on”) features, but the company is also collaborating with global automakers to apply the same model to higher levels of automation, including Level 3 (“Hands-off, eyes-off”) driving functions, letting drivers hand over responsibility to the AI driver in certain conditions.

The AI Foundation Model can efficiently support advanced driver assistance features like automatic emergency braking, speed control, and lane positioning, allowing drivers to take their hands off the wheel or be supported by system-initiated maneuvers, though they remain responsible for monitoring the driving task.

The AI software runs on Nvidia’s Drive hardware and is compatible with various sensor and compute architectures and so can be integrated into automakers’ current and future vehicle programs. As these vehicles operate in open, unconstrained environments, the AI learns from millions of real-world scenarios, including edge cases, which strengthen its capabilities and support Verification & Validation activities.

This ultimately leads to supporting Level 4 and Level 5 autonomy, updated over-the-air, allowing OEMs to gradually advance toward higher levels of autonomy says Kendall.

“Launching our US testing program in California deepens our collaboration with key partners like Microsoft, Nvidia and Uber,” said Kaity Fischer, VP of Commercial and Operations at Wayve. “Their support in cloud computing, silicon, and mobility services will accelerate the creation of a global ecosystem that will bring our AI-driving technology to automotive partners.”

The Wayve AI software enables car makers to provide a full range of driver assistance and automation features, replacing rule-based systems and geofenced routes with an end-to-end AI system that learns from data to drive any vehicle anywhere.

The company, launched in Cambridge in 2017, is backed by over $1.3bn from Softbank, Nvidia, Microsoft, Uber and Eclipse Ventures.

Further global expansion plans will be announced in the coming year.

www.wayve.ai

Leave a comment

Your email address will not be published. Required fields are marked *