| CPC G08G 1/164 (2013.01) [G06N 5/046 (2013.01); G06N 20/00 (2019.01); G08G 1/005 (2013.01); G08G 1/0112 (2013.01); G08G 1/0116 (2013.01); G08G 1/0129 (2013.01); G08G 1/0137 (2013.01); G08G 1/0141 (2013.01); G08G 1/017 (2013.01); G08G 1/052 (2013.01); G08G 1/093 (2013.01); G08G 1/096716 (2013.01); G08G 1/096725 (2013.01); G08G 1/123 (2013.01); G08G 1/163 (2013.01); G08G 1/166 (2013.01); G08G 1/0175 (2013.01)] | 20 Claims |

|
1. A method comprising:
using a visible-light camera located in a vicinity of a crosswalk that crosses a road to monitor an area in and nearby the crosswalk, the visible-light camera capturing a video of a vulnerable roadway user who is in or nearby the crosswalk;
determining a location and a movement direction of the vulnerable roadway user based on the video;
applying the location and the movement direction to a machine learning model running in equipment in the vicinity of the crosswalk to make a prediction that the vulnerable roadway users is about to enter the road in or nearby the crosswalk, the prediction being made prior to the vulnerable roadway user entering the road in or nearby the crosswalk,
wherein the machine learning model has been trained using motion data generated in the vicinity of the crosswalk, the motion data comprising at least one of location, speed, acceleration, and heading; and
before the vulnerable roadway user has entered the road in or nearby the crosswalk, causing a road sign to activate a light based on the prediction.
|