Apple gives us a look into its AI for its self-driving car project
Apple Inc has, in an online post, made public its research on how self-driving cars can better sport cyclists and pedestrains while using fewer sensors. The paper has been authored by Yin Zhou and Oncel Tuzel and submitted on Nov. 17 to independent online journal arXiv.
This is now a significant development because Apple's widel famed secrecy around future products which acted as a roadblock among artificial intelligence and machine learning researchers. The scientists have now proposed a new software approach which is called "VoxelNet" for helping computers detect three-dimensional objects.
Apple has declined to comment.
Apple in July had published the Apple Machine Learning Journal for its researchers. Their work seldom appears outside the journal, which has not published any research on self-driving cars.
Self-driving cars primarily use a combination of normal two-dimensional cameras and depth-sensing "LiDAR" units to recognize the world around them. While the units supply depth info, the low resolution makes it difficult to detect small, far off objects without the help of a normal camera linked to it in real time.
With the new software however, Apple researchers stated that they recieved "highly encouraging results" when it came to spotting pedestrains and cyclists on the roads with just LiDAR data. They also stated that they beat other approaches for detecting three-dimensional objects which also uses LiDAR. The experiments were computer simulations and did not involve road tests.
Though Chief Executive Tim COok had called self-driving cars "the mother of all AI projects," Apple had given few hints about the nature of its self-driving car ambitious.
Apple had told federal regulators last December that it was excited about the technology and had asked regulators not to restrict testing of the technology.
In April, Apple filed a self-driving car testing plan with California regulators
This is now a significant development because Apple's widel famed secrecy around future products which acted as a roadblock among artificial intelligence and machine learning researchers. The scientists have now proposed a new software approach which is called "VoxelNet" for helping computers detect three-dimensional objects.
Apple has declined to comment.
Apple in July had published the Apple Machine Learning Journal for its researchers. Their work seldom appears outside the journal, which has not published any research on self-driving cars.
Self-driving cars primarily use a combination of normal two-dimensional cameras and depth-sensing "LiDAR" units to recognize the world around them. While the units supply depth info, the low resolution makes it difficult to detect small, far off objects without the help of a normal camera linked to it in real time.
With the new software however, Apple researchers stated that they recieved "highly encouraging results" when it came to spotting pedestrains and cyclists on the roads with just LiDAR data. They also stated that they beat other approaches for detecting three-dimensional objects which also uses LiDAR. The experiments were computer simulations and did not involve road tests.
Though Chief Executive Tim COok had called self-driving cars "the mother of all AI projects," Apple had given few hints about the nature of its self-driving car ambitious.
Apple had told federal regulators last December that it was excited about the technology and had asked regulators not to restrict testing of the technology.
In April, Apple filed a self-driving car testing plan with California regulators
Comments
Post a Comment