After scaling back its ambitious car plans in 2016, Apple spent this year quietly immersing itself in the field of self-driving cars, and a new patent shows the company’s thinking of more efficient ways for autonomous vehicles to get around.

The patent application was published on Thursday and it has a simple title: . The application, as CNBC notes, lays out how Apple thinks it can make autonomous driving navigation more efficient. In part, this would be accomplished by “reducing the need to constantly remake detailed maps,” .

Here’s more from CNBC:

The patent claims that many autonomous vehicle systems base their navigation on static information — like maps — and use sensors to identify real-time information on the elements that change from day to day, as a way of minimizing the intense computing power needed to drive a car.

Instead, Apple’s system would be able to direct the car “independently of any data received from any devices external to the vehicle, and any navigation data stored locally to the vehicle prior to any monitoring of navigation.” Apple’s technology proposes a computerized model for predicting routes using sensors and processors in the vehicle.

It’s an interesting take; many in the industry view high-definition mapping data as a crucial element to making autonomous cars work. Some conceptions of the system involve user input, , where drivers are given the option to select alternative routes, or switch between manual and autonomous driving.

Of course, this is just a single patent—but the fact Apple’s thinking about offering an option between manual and automated driving strikes me as interesting. , Apple’s leasing an old Chrysler proving grounds in Arizona to test self-driving tech. Without plans to build a car on the table, I think the logical step for Apple to take would be to make the underlying hardware and software for autonomous driving work. A less-costly, more efficient system that’s outlined here is certainly one way to accomplish that—if it works.

A spokesperson for Apple didn’t immediately respond to a request for comment. If we hear back, we’ll update the post.