Back to results
US 2018/0005050 A1Published application
Autonomous vehicle localization using image analysis and manipulation
AssigneeUber Technologies, Inc.
InventorsBrett Browning, Peter Hansen, Ethan Eade, David Prasser, David LaRose, Robert Zlot
Life of this patent
Estimating the term…
Abstract
A system to use submaps to control operation of a vehicle is disclosed. A storage system may be provided with a vehicle to store a collection of submaps that represent a geographic area where the vehicle may be driven. A programmatic interface may be provided to receive submaps and submap updates independently of other submaps.
Claims21
A method for performing localization for a vehicle in motion, the method being implemented by one or more processors and comprising:
- accessing a collection of imagelets for a geographic region that encompasses multiple predefined areas of a road network, each imagelet depicting at least one feature present in a corresponding area of the multiple predefined areas, and wherein the at least one feature of each imagelet is associated with reference location information that is based on a location where sensor data of the imagelet was previously captured;
- obtaining sensor data from one or more sensor devices of the vehicle, the sensor data including (i) image data captured by one or more cameras of the vehicle, and (ii) data that identifies an approximate location of the vehicle;
- selecting a set of one or more imagelet from the collection of imagelets based at least in part on the approximate location of the vehicle;
- making a determination that an object depicted by the image data matches a feature depicted by at least one of the imagelets;
- wherein making the determination includes applying a transformation on at least one of the image data or the one or more imagelets, based on a determination that a condition exists which affects an appearance of objects in at least the approximate location of the vehicle;
- and
- determining a location of the vehicle that is more granular than the approximate location based at least in part on the reference location information associated with the feature depicted by the at least one of the imagelets.
4 more elements
A computer system comprising:
- a memory to store a set of instructions;
- one or more processors to use the set of instructions to:
- access a collection of imagelets for a geographic region that encompasses multiple predefined areas of a road network, each imagelet depicting at least one feature present in a corresponding area of the multiple predefined areas, and wherein the at least one feature of each imagelet is associated with reference location information that is based on a location where sensor data of the imagelet was previously captured;
- obtain sensor data from one or more sensor devices of the vehicle, the sensor data including (i) image data captured by one or more cameras of the vehicle, and (ii) data that identifies an approximate location of the vehicle;
- select a set of one or more imagelet from the collection of imagelets based at least in part on the approximate location of the vehicle;
- make a determination that an object depicted by the image data matches a feature depicted by at least one of the imagelets;
- wherein making the determination includes applying a transformation on at least one of the image data or the one or more imagelets, based on a determination that a condition exists which affects an appearance of objects in at least the approximate location of the vehicle;
- and
- determine a location of the vehicle that is more granular than the approximate location based at least in part on the reference location information associated with the feature depicted by the at least one of the imagelets.
6 more elements
Dependencies are inferred from the claim text.