Earlier this year, Luminar hired Christoph Schroeder, a veteran of Bosch and Daimler, to serve as its vice president of software. At Bosch, he was part of a team that pioneered radar technology that has become a staple of driver-assist systems, and he helped oversee production as radar transitioned from hundreds of prototypes and scaled toward hundreds of thousands of units.
“The prototype is not the problem,” Schroeder said. “The problem is making it work always. For a prototype, maybe you need 10 engineers. For the other, you need 100, not just to implement, but to test and document and make sure there is traceability, and to know why we do things. This will be out there for 15 years, and each car will have 200,000 miles on them. Every shortcut that you take will be exposed.”
In its efforts to test and document, Luminar has 150 quality-control checkpoints set up throughout its manufacturing process in Orlando, which combines space for manufacturing and high-end research.
Its units are built from the chip level up. Subassembly lines for components such as laser transceivers and optical beds feed into a final assembly line. Once the units are complete, they’re calibrated in one corner of the production floor, where white and gray tiles hang against a wall. The calibration ensures the lidar units provide returns with cameralike resolution, which is one reason why automakers say Luminar’s lidar stands out in a crowded marketplace.
Out the back door of the facility, there are targets in a tree line across a field. Luminar uses those to ensure each lidar unit reaches its stated range of 200 meters. That’s for today’s development model; it will increase to 250 meters for the Iris.
“The most important thing is that the product we’re shipping off the loading dock today meets the needs of the industry,” said Jason Eichenholz, the company’s chief technology officer. “The scale of what we’re doing today blows people’s minds. Our competition is trying to figure out how to catch up to where we were two years ago.”