The four-door Fusion sedan has long served as Ford's car of choice for testing autonomous driving technology, but you'd be hard-pressed to identify the radar, cameras, and other sensors on the latest self-driving Fusion.

Unveiled this week in a post on Medium, it has upgraded sensors that blend in with its styling. The only noticeable protrusions are two camera arrays that look like ski racks and two hockey-puck-sized LiDAR sensors on the front pillars that vaguely resemble police spotlights.

The net result is an ordinary-looking car, a testament to automotive normality made possible by advances in sensor technology, but also a byproduct of the fact that the focus of autonomous driving research isn't on the sensors, but rather on the computers inside the car.

On the latest Fusion, those computers process more than one terabyte of data per hour and run millions of pieces of code. Much of that computing power is dedicated to comparing what the sensors see outside the car to a pre-made 3D map of the surrounding environment, so the Fusion knows where traffic lights, crosswalks, and intersections are before it reaches them.

"When out on the road, the virtual driver uses its LiDAR, radar, and camera sensors to continuously scan the area around the car and compare—or mediate—what it sees against the 3D map," Ford Chief Program Engineer Chris Brewer wrote on Medium. Perfecting this method of "mediated perception" has been one of Ford's key goals for the past year—the company has invested in mapping startups and sensor manufacturers in addition to conducting its own research.

The ultimate goal with all the current research, Brewer wrote, is still to roll out a fleet of self-driving taxis by 2021, a project that the automaker announced in August. Unlike Uber's similar self-driving experiment, which encountered a major setback in California this month, Ford's won't rely on human drivers at all. The cars will even lack steering wheels and gas and brake pedals.