Earlier this week, Apple's (AAPL -2.98%) self-driving car prototype was again spotted on public roads in California. However, there were some notable changes this time around: Apple has significantly increased the number of sensors that are mounted on the roof. There are now no less than six lidar sensors mounted on top of a massive array that is affixed to the top of the vehicle, supplemented by numerous cameras and radar sensors. The sensor suite looks similar to prototypes that were spotted in August.
MacCallister Higgins, the co-founder of Voyage, spotted the prototype and shared the brief clip on social media. Voyage is a start-up that is developing autonomous taxis, so Higgins knows a thing or two here.
Going to need more than 140 characters to go over 🍎's Project Titan. I call it "The Thing" pic.twitter.com/sLDJd7iYSa— MacCallister Higgins (@macjshiggins) October 17, 2017
Here's another recent shot shared by a different user on social media.
I saw one of these a few weeks ago pull up to an Apple shuttle stop-sit there for a few then drove off. pic.twitter.com/gUudZY1TIA— idiggapple (@idiggapple) October 18, 2017
Compare that to the prototypes that were seen cruising around in April and May. For example, Bloomberg reporter Alex Webb saw one in late April.
The sheer size of the sensor array suggests that all of the computing hardware necessary to process the autonomous driving tasks is located within the array. That's a markedly different approach compared to other companies pursuing autonomous driving, which often put the computing hardware in the trunk where there's plenty of room. The inevitable question is why is Apple putting the compute stack inside the roof mounted array?
Does "Titan" refer to the size of the array?
There are two possible explanations. The first being that by developing a self-contained unit, Apple can more easily swap out prototypes during testing, potentially in order to test the system on different types of vehicles.
The far more interesting and speculative idea that's now floating around is that Apple intends to develop a modular array that could be retrofitted onto vehicles as an aftermarket upgrade that enables autonomous driving, similar to what Comma.ai had initially hoped to do before regulators nixed that idea (Comma.ai has since pivoted to selling a tool that allows people to hack into their vehicles). The chances of Apple actually going this route for its final product are somewhere between slim and none.
On a technical level, it's unlikely that such a modular upgrade would even work. Autonomous driving systems require deep integration with a vehicle's core driving functions. The systems aren't exactly comparable, but as another data point Tesla's Autopilot system can't be retrofitted because the procedure would be so invasive that the costs (mostly labor) would exceed buying a new vehicle, according to Elon Musk.
Should mention that retrofitting to full self-driving hardware is very difficult. Cost delta is more than buying a new car. Wish it weren't.— Elon Musk (@elonmusk) October 20, 2016
It would also be incredibly uncharacteristic for Apple to market an autonomous driving system as an aftermarket upgrade for existing vehicles. Not only would the company be compromising on hardware/software integration, the aesthetics -- something Apple cares deeply about -- would be atrocious. Not even Chief Design Officer Jony Ive could design a massive roof array that looks good on such a diverse set of vehicles. Ive is on record calling many modern car designs "shocking," so would he really want to mount one of his creations on top?
Apple CEO Tim Cook has finally publicly acknowledged that the company is "focusing on autonomous systems," which is also likely contributing to rising R&D expenditures. No one really knows what route Apple will take to market, and most of the choices are un-Apple-like. Perhaps Apple doesn't even know yet, but that's not stopping the Mac maker from moving forward on development.