How Microsoft just got into LiDAR Free Self-Driving AI

When I was doing my research on the AV sector, one startup stood out to me as working significantly on Deep learning of autonomous vehicles and that company was Wayve. So when I woke up this morning and found Microsoft has invested heavily in them, I wasn't exactly surprised.

Microsoft Research have noticeably improved since the pandemic and put Microsoft on the map in AI research.

I like Wayve's story a lot. Founded in London in 2017, Wayve’s team of machine learning scientists and roboticists are trying to pioneer an artificial intelligence-first approach to autonomous driving. Alex Kendall, the New Zealander who co-founded Wayve, said that their approach to AVs is "quite contrarian". I had recently written about this as well.

Wayve does not rely on LiDAR sensors but on Deep learning itself. Projects like Google’s Waymo, Uber, Cruise and Aurora are developing autonomous vehicles by throwing engineers at the problem. Wayve is throwing AI and deep learning at the problem.

Based in London, Wayve uses end-to-end deep learning to develop artificial intelligence capable of complex driving. Microsoft just became their biggest backer.

Autonomous driving start-up Wayve bags $200 million from Microsoft, Virgin and Baillie Gifford

Their current valuation after the funding should be close to $2 billion now. So far then they have raised $258 million. LiDAR-free AV systems are just so much less expensive to scale and implement. As a Moonshot then, it's just so valuable potentially. I'm glad that Microsoft understood the opportunity here.

Do not underestimate Deep learning in solving the autonomous driving bottlenecks. Wayve has made some pretty bold and crazy claims in the past. Including a “world first” in demonstrating that a car working on their machine-learning platform can drive on roads it’s never seen before during training, and without an HD map of its environment.

Wayve is an alumni of the University of Cambridge. Their cars learn to drive from data with machine learning. Every time a safety driver intervenes and takes over, the car learns to drive better. Over the years their system is particular, they don’t tell the car how to drive, rather it learns to drive from experience, example and feedback, just like a human. This is more safe and scalable than any other approach today some academics and engineers in the field believe.

It's a radical approach if you think about it. The traditional approach used by all our competitors relies on HD-maps, expensive sensor suites and hand-coded rules that tell the car how to drive. We have built a system that learns end-to-end with machine learning. It is the first in the world to drive on urban roads it has never been on before. It uses compute/sensors which cost less than 10% of competitors.

Wayve must have made some serious headway with this unusual approach to AV, for Microsoft to get so involved. What do you think?

Wayve has partnered with companies operating fleets of delivery trucks, including grocers Ocado and Asda, to collect driving data to improve Wayve’s AI technology. Wayve most recently entered a?partnership ?with parcel delivery company DPD.

Wayve believes that deep learning has an important role to play in autonomous driving. Deep learning is an area of AI that attempts to mimic the activity in layers of neurons in the brain to learn how to recognize complex patterns in data.

Wayve has chosen to license its autonomous driving technology to commercial fleets instead of trying to manufacture its own full self-driving vehicles, which are yet to go on sale to the public. With such a big early round, Microsoft will have quite a share in Wayve's future it now appears.

If you want to follow my articles on AI subscribe to AiSupremacy, my main Newsletter for daily topics here. I need community support to be able to continue writing here.

https://aisupremacy.substack.com/subscribe