Apple Wins a Second 'Project Titan' Patent Today covering an advanced LiDAR-Centric 3D Object Detection System
Yesterday Patently Apple posted a report titled "A Credible Report is claiming that Sources have seen Apple's Next-Gen Vehicle Battery Technology that is described as 'Radical.' A secondary part of the Reuters report that we didn't cover was its mention of Apple's advanced work on LiDAR sensors.
Reuters reported that "Apple has decided to tap outside partners for elements of the system, including lidar sensors, which help self-driving cars get a three-dimensional view of the road, two people familiar with the company’s plans said.
Apple’s car might feature multiple lidar sensors for scanning different distances, another person said. Some sensors could be derived from Apple’s internally developed lidar units, that person said. Apple’s iPhone 12 Pro and iPad Pro models released this year both feature lidar sensors."
Today the U.S. Patent and Trademark Office officially granted Apple their second "Project Titan" patent of the day titled "Three-dimensional object detection."
Apple's invention covers vehicle navigation (e.g., manual, semi-autonomous, fully autonomous) system that requires the accurate detection of objects (e.g., other vehicles, pedestrians, buildings, etc.) along a navigated pathway. Vehicles can use sensors (e.g., embedded vehicle sensors, external sensors that communicate information wirelessly to the vehicle, etc.) to detect information associated with the objects to enable the proper vehicle navigation.
Apple's invention provides a method and system for detecting objects (e.g., three-dimensional objects) in a three-dimensional space using sensor information (i.e., remote sensing information) and for determining information associated with the detected objects using the sensor information.
The method and system utilizes a detection algorithm (e.g., a 3D detection algorithm) to directly model the orientation of the objects into the three-dimensional space.
The sensor information can include but is not limited to LIDAR point clouds (e.g., 3D LIDAR point clouds). The information associated with the detected objects can include but is not limited to locations (i.e., x, y, z coordinates), shape/size (i.e., width and height measurements), and rotation orientation.
Apple's invention also provides a system that performs an end-to-end deep neural network (DNN) mechanism or other machine learning system to accurately detect objects and to accurately determine rotation information from detected sensor information including but not limited to point cloud data. The machine learning system can be used for multi-modal active learning, annotation intelligence, and autonomy perception.
Apple's patent FIG. 9 below is an illustration that shows a vehicle. The vehicle (#800) includes an image sensor #900 and a range sensor #902. Information from the image sensor and the range sensor are provided to the automated control system #808, which is operable to process and store information associated with detecting three-dimensional objects.
Apple's patent FIG. 2A above shows range points in an environment in a top-down view; FIG. 7 is an illustration that shows object detector #722.
Review Apple's granted patent 10,872,228 for finer details.
Velodyne Lidar and Luminar Technologies stock prices got a healthy boost from yesterday's report about Apple Car 2024 using advanced LiDAR systems like the one outlined today. The stocks rose 13.5% and 9% respectfully.