公开(公告)号：US20230221942A1

公开(公告)日：2023-07-13

申请号：US17647524

申请日：2022-01-10

Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLC

Inventor： Wende Zhang ,  Esther Anderson ,  Mark T Gaylord ,  Robert A Dziurda

IPC: G06F8/65 ,  G06F8/71 ,  G06K9/62

CPC classification number: G06F8/65 ,  G06F8/71 ,  G06K9/6277 ,  G06K9/6257

Abstract: Vehicles and related systems and methods are provided for classifying detected objects in a location-dependent manner using localized models in a federated learning environment. A method involves obtaining sensor data for a detected object external to the vehicle from a sensor of a vehicle, obtaining location data associated with the detected object, obtaining a local classification model associated with an object type, assigning the object type to the detected object based on an output by the local classification model as a function of the sensor data and the location data using the local classification model, and initiating an action at the vehicle responsive to assigning the object type to the detected object.

公开(公告)号：US11853741B2

公开(公告)日：2023-12-26

申请号：US17647524

申请日：2022-01-10

Applicant: GM GLOBAL TECHNOLOGY OPERATIONS LLC

Inventor： Wende Zhang ,  Esther Anderson ,  Mark T Gaylord ,  Robert A Dziurda

IPC: G06F8/65 ,  G06F8/71 ,  G06F18/214 ,  G06F18/2415 ,  G06T7/70 ,  G06V10/00 ,  G06V20/00

CPC classification number: G06F8/65 ,  G06F8/71 ,  G06F18/2148 ,  G06F18/2415 ,  G06T7/70 ,  G06V10/00 ,  G06V20/00 ,  G06T2207/20081 ,  G06T2207/20084 ,  G06T2207/30252

Abstract: Vehicles and related systems and methods are provided for classifying detected objects in a location-dependent manner using localized models in a federated learning environment. A method involves obtaining sensor data for a detected object external to the vehicle from a sensor of a vehicle, obtaining location data associated with the detected object, obtaining a local classification model associated with an object type, assigning the object type to the detected object based on an output by the local classification model as a function of the sensor data and the location data using the local classification model, and initiating an action at the vehicle responsive to assigning the object type to the detected object.