| CPC B60W 60/0011 (2020.02) [B60W 40/04 (2013.01); B60W 50/0097 (2013.01); B60W 60/00274 (2020.02); G06V 20/58 (2022.01); H04W 4/40 (2018.02); B60W 2420/403 (2013.01); B60W 2554/4045 (2020.02); B60W 2556/50 (2020.02); B60W 2556/65 (2020.02); G06V 2201/08 (2022.01)] | 20 Claims |

|
1. A method comprising:
obtaining a driving scenario, wherein the driving scenario comprises a representation of a road segment and a driving destination of an autonomous vehicle driving in the road segment, wherein the driving scenario comprises a second vehicle in the road segment;
selecting at least two different driving models that are usable for predicting movement patterns of the second vehicle, wherein the at least two different driving models are selected from a predefined plurality of driving models, wherein the predefined plurality of driving models is known prior to said obtaining the driving scenario;
predicting at least two movement patterns of the second vehicle within the driving scenario according to the at least two different driving models, respectively;
for each one of the at least two movement patterns of the second vehicle, generating a respective movement pattern for the autonomous vehicle, thereby generating at least two movement patterns for the autonomous vehicle;
training an autonomous driving model based on the at least two movement patterns generated for the autonomous vehicle, thereby obtaining a trained autonomous driving model; and
providing the trained autonomous driving model to a second autonomous vehicle which is configured to use the trained autonomous driving model for autonomous driving.
|