So far, Apple's work on self-driving cars has been kept under wraps — at least compared to other companies working toward the same goal. But now, two scientists at the company have published some of their research progress .
It's rare for a company like Apple to show off some of its research or explain its products before the initial launch date, but autonomous vehicles have important specialties to consider. For example, to test these vehicles in public requires acquiring authorization from governments , so it's important for the company to allow access to its files. Added to this, as mashable.com points out , are the complications of testing a vehicle in public without anyone noticing.
Engineers from the Cupertino company, Yin Zhou and Oncel Tuzel , have published details in the independent journal arXiv of a new software technique that would help autonomous vehicles distinguish pedestrians uganda phone number cyclists . The technique, called VoxelNet , works without the need for a camera. The scientists therefore argue that this new method could be more effective than those that use a camera system, in addition to LiDAR (Light Detection and Ranging) , a device that allows the distance to an object to be determined using a beam of light. Until now, this type of expensive system was the standard in the industry, using cameras to determine small or distant objects, such as pedestrians or cyclists.
VoxelNet, named after the voxel , which is the value of a point on a three-dimensional grid, allows LiDAR alone to be used to detect distant objects. Scientists have already tested the software, although they have only tried it in computer simulations . Because of this, Apple will need to use VoxelNet on the road before it can finally remove cameras from its autonomous vehicles. These early results have been described as “very encouraging” , so that may happen relatively soon.