US 11,834,067 B1
Comfort scales for autonomous driving
Keith Hutchings, San Francisco, CA (US); Ilmo Konstantin van der Löwe, Fremont, CA (US); and Taylor Boyce Bixby, Berkeley, CA (US)
Assigned to Waymo LLC, Mountain View, CA (US)
Filed by Waymo LLC, Mountain View, CA (US)
Filed on Jul. 30, 2021, as Appl. No. 17/390,261.
Int. Cl. B60W 60/00 (2020.01); G08G 1/01 (2006.01); G06V 20/56 (2022.01); G06F 18/214 (2023.01)
CPC B60W 60/001 (2020.02) [G06F 18/214 (2023.01); G06V 20/56 (2022.01); G08G 1/0125 (2013.01); B60W 2554/4049 (2020.02)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method comprising:
receiving data representing a traffic encounter between a vehicle and a pedestrian;
receiving a plurality of comfort scale ratings of the traffic encounter assigned by a rater belonging to a first rater pool;
generating an input element for a rating transformation model configured to predict how comfort scale ratings assigned by a particular rater pool would have been assigned by a representative rater belonging to a reference rater pool;
performing an inference pass over the rating transformation model using the input element to obtain a plurality of transformed comfort scale ratings for the reference rater pool;
training a driving control model using the transformed comfort scale ratings; and
controlling an autonomously driven vehicle (ADV) based on the driving control model trained on the plurality of transformed comfort scale ratings for the reference rater pool.