Providing actionable uncertainties in autonomous vehicles

A computing system, comprising:

• one or more first machine-learned models that associated with detect objects external to an autonomous vehicle based at least in part on sensor data;
• one or more second machine-learned models that generate predictions in association with objects external to the autonomous vehicle;
• one or more third machine-learned models that generate motion plans in association with the autonomous vehicle;
• one or more processors;
• and
• one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
• generating, as an output of the one or more first machine-learned models, data indicative of one or more detected objects external to the autonomous vehicle;
• generating data indicative of at least one uncertainty associated with the one or more detected objects;
• inputting the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty associated with the one or more detected objects to the one or more second machine-learned models;
• generating, as an output of the one or more second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects, the at least one prediction based at least in part on the one or more detected objects and the at least one uncertainty;
• inputting the data indicative of the at least one uncertainty associated with the one or more detected objects and the data indicative of the at least one prediction associated with the one or more detected obj ects to the one or more third machine-learned models;
• and
• generating, as an output of the one or more third machine-learned models, data indicative of at least one motion plan based on the one or more detected objects;
• wherein the data indicative of the at least one motion plan is generated by the one or more third machine-learned models based at least in part on the data indicative of the at least one uncertainty associated with one or more detected objects and the data indicative of the at least one prediction associated with the one or more detected objects.

An autonomous vehicle, comprising:

• one or more processors;
• and
• one or more non-transitory computer-readable media that collectively store instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, the operations comprising:
• inputting sensor data to one or more first machine-learned models that detect objects external to an autonomous vehicle;
• obtaining, as an output of the one or more first machine-learned models, data indicative of one or more detected objects external to the autonomous vehicle;
• determining data indicative of at least one uncertainty associated with the one or more detected objects at a current cycle, wherein the data indicative of the at least one uncertainty is determined based on the data indicative of the one or more detected objects and history data associated with the one or more detected objects at a previous cycle;
• inputting the data indicative of the one or more detected objects and the data indicative of the at least one uncertainty associated with the one or more detected objects to one or more second machine-learned models that generate predictions in association with objects external to the autonomous vehicle;
• and
• obtaining, as an output of the one or more second machine-learned models, data indicative of at least one prediction associated with the one or more detected objects, the at least one prediction based at least in part on the one or more detected objects and the at least one uncertainty.