Unified boundary machine learning model for autonomous vehicles

An autonomous vehicle control system for an autonomous vehicle, comprising:

• one or more processors;
• and
• memory storing instructions that, when executed by the one or more processors, cause the autonomous vehicle control system to:
• receive perception data from at least one perception sensor positioned to sense a roadway upon which the autonomous vehicle is disposed;
• generate a plurality of perceived boundaries for the roadway by processing the perception data using a trained machine learning model that integrates detection of boundaries with detection of attributes for at least a subset of the plurality of perceived boundaries, wherein at least one perceived boundary of the plurality of perceived boundaries that is generated using the trained machine learning model is defined by a plurality of spaced apart construction elements that are detected and linked together into a virtual boundary, and wherein the trained machine learning model is configured to receive the perception data;
• output, by the trained machine learning model and in response to receiving the perception data, each of the plurality of perceived boundaries;
• for each of the plurality of perceived boundaries for which an associated attribute is detected, output, by the trained machine learning model and in response to receiving the perception data, the associated attribute therefor;
• determine, using the attribute output from the trained machine learning model for a first perceived boundary output from the trained machine learning model, a drivable or non-drivable region of the roadway;
• determine one or more active lanes using the plurality of perceived boundaries;
• and
• control the autonomous vehicle using the one or more active lanes and the plurality of perceived boundaries and the attributes thereof.

A method of operating an autonomous vehicle with an autonomous vehicle control system, comprising:

• receiving perception data from at least one perception sensor positioned to sense a roadway upon which the autonomous vehicle is disposed;
• generating a plurality of perceived boundaries for the roadway by processing the perception data using a trained machine learning model that integrates detection of boundaries with detection of attributes for at least a subset of the plurality of perceived boundaries, wherein at least one perceived boundary of the plurality of perceived boundaries that is generated using the trained machine learning model is defined by a plurality of spaced apart construction elements that are detected and linked together into a virtual boundary, and wherein the trained machine learning model is configured to receive the perception data;
• outputting, by the trained machine learning model and in response to receiving the perception data, each of the plurality of perceived boundaries;
• for each of the plurality of perceived boundaries for which an associated attribute is detected, outputting, by the trained machine learning model and in response to receiving the perception data, the associated attribute therefor;
• determining, using the attribute output from the trained machine learning model for a first perceived boundary output from the trained machine learning model, a drivable or non-drivable region of the roadway;
• determining one or more active lanes using the plurality of perceived boundaries;
• and
• controlling the autonomous vehicle using the one or more active lanes and the plurality of perceived boundaries.