Some Information About Their Self-Driving Car Research Has Been Disclosed By Apple Scientists For The First Time


11/22/2017



It is for the first time that information about what researchers and computer scientists at Apple Inc. are working on with respect to self-driving cars have been revealed by the company itself. In a first publicly disclosed paper on autonomous vehicles being researched by the company revealed online, it revealed the work that they are doing on how self-driving cars would be able to better spot cyclists and pedestrians while they make use of fewer sensors.
 
With respect to artificial intelligence and machine learning researchers, the corporate secrecy around future products that have been created by Apple, has been viewed as a major drawback. The information, which has thus assumed significance,  was revealed in a paper that was prepared by Yin Zhou and Oncel Tuzel, and was submitted on Nov. 17 to the independent online journal arXiv.
 
To enable computers to be able to easily and more effectively detect three-dimensional objects, the researchers at Apple have proposed a new software approach that has been called "VoxelNet".
 
There were no comments available from Apple.
 
Researches and academicians typically freely share their work with their peers and researchers from other organizations. The Apple Machine Learning Journal was started by Apple in July for its researchers to share their work after yielding to the practice of sharing research developments. However, information about their research by Apple’s scientists have not yet been published elsewhere and no research work on self-driving cars have so far been published in the Apple’s’ journal.
 
In order to identify and recognize the world around them, both two dimensional cameras and depth-sensing “LiDAR” units are utilized by self-driving cars. There is need for a normal camera linked to the units in real time because even though units supply depth of information, the images are generally shot in low resolution that makes it hard for computers to identify and detect small and faraway objects.
 
But using just LiDAR data, Apple researchers have been able to obtain “highly encouraging results” in spotting pedestrians and cyclists with the help of the new software. Additionally, other generally used approaches for the detection of three-dimensional objects which make use of only LiDAR were beaten by the Apple researchers, they wrote in the article. All of the experimental trials were however conducted on computer and no road tests were involved.
 
Very little information about the type of self-driving cars that Apple is developing even though self driving cars have been termed as the “the mother of all AI projects,” by Chief Executive Tim Cook.
 
Apple had asked the U.S. federal regulators not to restrict the self driving technology, last December, because it was excited about the technology.
 
A self-driving car testing plan was applied for by Apple with the California regulators in April.
 
(Source:www.reuters.com)