Back to results
US 12,736,661 B1Granted
Labelling object data with radar signatures
AssigneeZoox, Inc.
InventorArvind Srivastav
Life of this patent
Estimating the term…
Abstract
Sensors may be used to detect objects in an environment. In an example, a vehicle may include a radar sensor and one or more additional sensors. The additional sensors generate sensor data used to detect objects in the environment and determine attributes associated with the detected objects. Radar data from the radar sensor is compared to the attributes of the detected object to identify radar returns that are associated with the object. Object data is generated that associates velocity information for the identified radar returns with the object.
Claims19
A vehicle comprising:
- a lidar sensor associated with the vehicle;
- a radar sensor associated with the vehicle;
- one or more processors;
- and
- memory storing non-transitory computer-readable instructions that, when executed, configure the one or more processors to perform operations comprising:
- receiving lidar data from the lidar sensor;
- determining, based at least in part on the lidar data and at the exclusion of radar data, an attribute of an object in an environment, wherein the determining the attribute of the object further comprises determining three-dimensional coordinates of the object in the environment by determining instances of the three-dimensional coordinates of the object at a first frequency;
- receiving radar data from the radar sensor, the radar data including Doppler velocity information and one or more of range information, azimuth information, or elevation information, wherein instances of the radar data are received at a second frequency higher than the first frequency;
- comparing the one or more of the range information, the azimuth information, or the elevation information to the attribute of the object;
- determining, based at least in part on the comparing, a subset of radar points of the radar data associated with the object;
- generating a two-dimensional projection associated with the subset of the radar points, the two-dimensional projection comprising Doppler velocities of individual points of the radar points relative to depths of the individual points of the radar points;
- generating object data that associates the two-dimensional projection with the object, the generating the object data comprising associating the Doppler velocities of the radar points associated with multiple instances of the radar data generated at different times with a single instance of the object data based on multiple instances of the radar data associated with a single instance of the three-dimensional coordinates of the object;
- sending the object data to a machine-learned model, the machine-learned model being configured to output an expected motion of the object;
- receiving, as an output of the machine-learned model, the expected motion of the object;
- and
- controlling the vehicle based at least in part on the expected motion of the object.
13 more elements
A method comprising:
- determining, from first sensor data, at least one of an estimated velocity or an estimated yaw rate of an object in an environment;
- determining, from the first sensor data, three-dimensional coordinates of the object in the environment by determining instances of the three-dimensional coordinates of the object at a first frequency;
- receiving, from a radar sensor, radar data associated with the environment, the radar data comprising radar points having associated Doppler velocities, wherein instances of the radar data are received at a second frequency higher than the first frequency;
- comparing the radar data to at least one of the estimated velocity or the estimated yaw rate;
- determining, based at least in part on the comparing of the radar data to the at least one of the estimated velocity or the estimated yaw rate, a subset of the radar points associated with the object;
- generating a two-dimensional projection associated with the subset of the radar points, the two-dimensional projection comprising Doppler velocities of individual points of the radar points relative to depths of the individual points of the radar points;
- generating object data that associates the two-dimensional projection Doppler velocities of the subset of the radar points with the object, the generating the object data comprising associating the Doppler velocities of the radar points associated with multiple instances of the radar data generated at different times with a single instance of the object data based on multiple instances of the radar data associated with a single instance of the three-dimensional coordinates of the object;
- determining, based at least in part on the object data, an expected motion of the object;
- and
- controlling a vehicle based at least in part on the expected motion of the object.
7 more elements
Non-transitory computer-readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
- determining, from first sensor data, at least one of an estimated velocity or an estimated yaw rate of an object in an environment;
- determining, from the first sensor data, three-dimensional coordinates of the object in the environment by determining instances of the three-dimensional coordinates of the object at a first frequency;
- receiving, from a radar sensor, radar data associated with the environment, the radar data comprising radar points having associated Doppler velocities, wherein instances of the radar data are received at a second frequency higher than the first frequency;
- comparing the radar data to at least one of the estimated velocity or the estimated yaw rate;
- determining, based at least in part on the comparing of the radar data to the at least one of the estimated velocity or the estimated yaw rate, a subset of the radar points associated with the object;
- generating a two-dimensional projection associated with the subset of the radar points, the two-dimensional projection comprising Doppler velocities of individual points of the radar points relative to depths of the individual points of the radar points;
- generating object data that associates the two-dimensional projection Doppler velocities of the subset of the radar points with the object, the generating the object data comprising associating the Doppler velocities of the radar points associated with multiple instances of the radar data generated at different times with a single instance of the object data based on multiple instances of the radar data associated with a single instance of the three-dimensional coordinates of the object;
- determining, based at least in part on the object data, an expected motion of the object;
- and
- controlling a vehicle based at least in part on the expected motion of the object.
7 more elements
Dependencies are inferred from the claim text.