With the acquisition of music app company Shazam on the horizon, it is easy to forget that tech giant Apple does more than devices and media. Like Google, Facebook, and other contemporaries, the company has been quietly funneling significant resources into artificial intelligence research. While it may be a surprise to many, the software and computer tech company apparently has an eye for the world of self-driving cars, and while they have largely remained secretive about the project, Apple has recently opened up about their progress.
The first inklings of Apple’s progress in the automotive world came in April when the company applied for a permit to test their self-driving vehicles in their home state of California. Only months later in June, Tim Cook, Apple’s current CEO, confirmed the company’s active projects in developing artificial intelligence software and indicated quite directly that it could one day be used for autonomous vehicles. However, the company’s biggest update on their progress came this week after Ruslan Salakhutdinov, Apple’s director of their artificial intelligence research and projects, gave a talk this past Friday.
Salakhutdinov treated an audience of around 200 artificial intelligence tech experts at the Neural Information Processing Systems conference to a talk about the company’s progress with machine learning. The director revealed that Apple had developed a sophisticated AI capable of analyzing data provided by both cameras and LIDAR. At its simplest, LIDAR is a detection method that uses a pulsed laser to measure distances. According to Salakhutdinov, Apple has succeeded in building an AI system that can detect pedestrians, cars, cyclists, and even monitor driving lane markers.
The advanced system could even maintain its tracking when raindrops obstructed the camera’s lense and, in a remarkable showing, could estimate the position of pedestrians obscured behind other stationary vehicles. Salakhutdinov also went into extensive detail about the software behind the map building capabilities needed by autonomous vehicles. Self-driving cars require detailed maps to track traffic lights, lanes, and obstacles, and based off Salakhutdinov’s presentation, Apple has at least two projects working towards this goal.
Even with these revelations from Salakhutdinov, the future of Apple self-driving cars remains unclear. While the talk revealed many smaller projects, how all of the demonstrated systems will come together into a consumer product was not specified. That said, the potential of Apple entering the autonomous vehicle market, given the sheer success of their more practical devices, could change a market otherwise dominated by Google and Tesla.