Addressing the various challenges that arise in autonomous driving is an incredibly complex task, but making an effort to get started means making sure you have quality data that is accurate and well annotated. This is where scale comes in, given the early recognition that the AV industry will need the annotation of vast swaths of data, including specialized lidar imaging. Now, co-founder and CEO Alex Wang tells me at TC Sessions: Mobility 2021 (additional Crunch subscription required) that it’s moving into the mapping with a new product coming later this month.
“Our role is constantly evolving,” Wang said, explaining how it works with its transportation industry partners, including Toyota, among many others. “You know, as we work with our customers, and we’ve solved a problem for them around labeling data and annotation data, you know, it turns out they come to us with other problems. We launched a product called Nucleus for that. A lot of our customers are thinking a lot about mapping, and how to deploy with more robust maps So we’re building a product, I’m going to announce it later this month, but we’re working with our customers to help solve that problem.”
Despite my upvoting, Wang wouldn’t provide any specifics, but he did go into more detail about the challenges of mapping, and what existing maps are lacking for companies working on integrating those with AV systems. , which includes other signals, such as sensor fusions and vehicle-to-infrastructure components.
“I think a big question for the overall location has been that historically, the industry has relied heavily on mapping – we relied heavily on very high-quality, high-definition maps,” he said. “The hard thing about the world is that sometimes these maps are wrong, and how do you deal with that? […] How do you deal with this kind of challenge of robustness, or updates. Even, if you think about it, Google Maps, which has the best mapping infrastructure in the world, by a wide margin, you know they don’t update fast enough for this. [human] Driver.”
Wang said the challenge is no different from the challenge that Scale has been actively solving for most of its existence, which is the data flywheel. With autonomous driving, it is extremely important to be able to quickly and accurately collect and annotate data, resulting in better collection and annotation of future data, and the assumptions the system is making about its environment. for greater reliability.
“Finding how to deal with the real-time nature of how the world changes is a really big, a huge component,” he said. While we still have to wait to see exactly what Scale has planned, it’s safe to assume that it’s going to build confidence in maps and use accuracy as a key component in everything they launch It’s about mapping.