Aeva’s lidar was destined to get a lot of attention, because, well, the company was founded by ex-Apple engineers. That means comes with a certain amount of built-in excitement, because maybe it will help us finally answer the question—what would Apple engineers do if we asked them to solve the big problem of automotive lidar?
The answer is that they would approach the problem like a bunch of Apple engineers.
Design and operating principles
Here’s the breakdown. Aeva says the lidar is solid state and claims a range of 200 meters. It’s slated to cost “a few hundred bucks” when produced at volume. The device is beautifully designed, so much so that I wouldn’t mind keeping on top of my car instead of hiding away in the front grill.
It differs from a conventional lidar in that it shoots out a continuous “wave” of light rather than individual pulses (a lot like Cyclops). Ars Technica explains that the system steadily changes the frequency of the wave, and when the wave bounces back, combines the incoming and outgoing wave. I’ll let Ars finish:
“Because the frequency is changing at a constant rate, there’s a frequency difference between the two beams that’s directly proportional to the distance that the returning beam traveled before bouncing back. Combining beams with two slightly different frequencies produces a beat frequency that gives a precise estimate of the frequency difference—and hence a precise distance measurement.”
Aeva founder Mina Rezk has said that this allows the sensor to capture a more detailed image of the car’s surroundings, as well as measure velocity. Knowing velocity allows it to flag objects that are in motion—a running child, for instance—and tell the vehicle’s operating system to pay attention.
If you think Aeva’s sensor sounds a lot like another lidar sensor that has received a lot of attention recently, you’re right. Blackmore’s lidar sensor also exploits frequency modulation. More importantly, both belong to the new class of frequency modulated continuous wave lidar (FMCW), which is said to reduce interference from other sensors and offer better performance during inclement weather. Lastly, both are part of a growing number of automotive lidar that can measure velocity.
According to what Aeva told Wired, they hope to beat out the various hardware options that capture velocity information by offering an integrated solution. Most lidar sensors for autonomous vehicles are just lidar sensors that designers still need to integrate with other sensors—radar, cameras, etc—to build a complete sensor payload. Aeva’s product is more of a full solution that includes lidar alongside all the other sensors that autonomous cars use.
Aeva’s sensor package comes fully calibrated and integrated out of the box–which could have secondary benefits for odometry. The sensor package even offers built-in computer vision tech to perform sensor fusion duties within the box itself. Reading the Aeva feature in The Verge, it’s clear that the company believes that offering an integrated product with basic fusion will appeal to companies that want to build their own self-driving vehicle system without cobbling together a bunch of components that were not explicitly designed to work with one another.
So, what did I mean when I said that Aeva approached automotive lidar like a bunch of Apple engineers? I mean they integrated a bunch of technologies into a single box with the hope that it would *just work* right out of the box. They also made that box look great. Whether that’s enough to help them take the market, only time will tell. Since Aeva hasn’t gotten the product to market yet, all we can do for now is speculate.