Modeling behavior based on observations of objects observed in a driving environment
AssigneeGoogle Inc.
InventorsDavid I. Ferguson, Jiajun Zhu, Dmitri A. Dolgov
Life of this patent
Estimating the term…
Abstract
A method and apparatus are provided for determining one or more object models used by an autonomous vehicle to predict the behavior of detected objects. The autonomous vehicle may collect and record object behavior using one or more sensors. The autonomous vehicle may then communicate the recorded object behavior to a server operative to determine the object models. The server may determine the object models according to a given object classification, a particular driving environment, or a combination thereof. The server may then communicate the object models to the autonomous vehicle for use in predicting the actions of detected objects and for responding accordingly.
Claims23
A system for determining behavior data used by an autonomous vehicle, the system comprising:
- a memory configured to store an object model;
- and
- one or more processors in communication with the memory, the one or more processors configured to:
- identify a type of driving environment and a geographic location of the driving environment;
- receive, from a monitoring source, object data associated with a detected object;
- analyze the object data to determine an object classification for the detected object, the object classification identifying a class of objects;
- determine whether an object model currently exists for the object classification, wherein the object model correlates the expected behavior of objects within the determined object classification with the identified type of driving environment and the identified geographic location;
- determine a new object model for the object classification based on the received object data when an object model does not currently exist for the object classification;
- update the currently-existing object model for the object classification based on the received object data when an object model currently exists for the object classification;
- and
- control the autonomous vehicle based on the object model.
A method for determining behavior data used by an autonomous vehicle, the method comprising:
- identifying a type of driving environment and a geographic location of the driving environment;
- receiving, with one or more processors, from a monitoring source, object data associated with a detected object;
- analyzing, with the one or more processors, the object data to determine an object classification for the detected object, the object classification identifying a class of objects;
- determining whether an object model currently exists for the expected behavior of objects within the determined object classification in the identified type of driving environment and at the identified geographic location;
- determining a new object model for the object classification based on the received object data when an object model does not currently exist for the object classification;
- updating the currently-existing object model for the object classification based on the received object data when an object model currently exists for the object classification;
- and
- controlling the autonomous vehicle based on the object model.
A system for determining behavior data used by an autonomous vehicle, the apparatus system comprising:
- a memory configured to store an object model;
- and
- one or more processors in communication with the memory, the one or more processors configured to:
- identify a type of driving environment and a geographic location of the driving environment;
- receive object data associated with a plurality of detected objects from a source that monitored the behavior of the plurality of detected objects in the identified type of driving environment and the identified geographic location;
- identify at least one object classification for each of the detected objects based on the received object data, each object classification identifying a class of objects
- determine, for each identified object classification, whether an object model currently exists for the expected behavior of objects within the identified object classification in the identified type of driving environment and at the identified geographic location;
- determine a new object model for an identified object classification based on the received object data when an object model does not currently exist for the object classification;
- and
- update the currently-existing object model for an identified object classification based on the received object data when an object model currently exists for the object classification;
- and
- control the autonomous vehicle based on the object model.
Dependencies are inferred from the claim text.