| CPC B60W 30/143 (2013.01) [B60W 40/04 (2013.01); B60W 50/14 (2013.01); B60W 2050/146 (2013.01); B60W 2555/60 (2020.02)] | 20 Claims |
|
1. A method comprising:
obtaining, via one or more sensors of a vehicle, vehicle sensor data pertaining to operation of the vehicle;
obtaining, via a transceiver, traffic light data with respect to a plurality of traffic lights along a path or roadway on which the vehicle is travelling, the traffic light data including:
with respect to a first traffic light along the path or roadway in which the vehicle is travelling, multiple first time intervals in which the first traffic light is expected to be green, along with multiple second time intervals in which the first traffic light is expected to be yellow or red, including multiple changes of the first traffic light between green and red; and
with respect to a second traffic light along the path or roadway in which the vehicle is travelling, multiple third time intervals in which the second traffic light is expected to be green, along with multiple fourth time intervals in which the first traffic light is expected to be yellow or red, including multiple changes of the first second traffic light between green and red;
determining, via a processor, a desired control of movement of the vehicle based on the vehicle sensor data, the traffic light data, and one or more optimization criteria pertaining to the vehicle, including a desired speed for the vehicle, such that the vehicle travels through the first traffic light during one of the multiple first time intervals in which the first traffic light is expected to be green, and such that the vehicle also travels through the second traffic light during one of the multiple third time intervals in which the second traffic light is expected to be green, while optimizing the one or more optimization criteria; and
taking a vehicle action in accordance with instructions provided by the processor based on the desired control of movement of the vehicle, wherein the step of taking the vehicle action comprises automatically controlling a speed of the vehicle to meet the desired speed, via the instructions provided by the processor and implementation of the instructions via one or more vehicle systems coupled to the processor;
wherein the desired control of movement of the vehicle is determined by the processor based on the vehicle sensor data using:
coarse-grain speed planning, via machine learning applied by the processor to the sensor data with respect to both the first traffic light and multiple consecutive traffic lights of the plurality of traffic lights, and including in generating a list or range of temporal-spatial locations of vehicle arrival with respect to the first traffic light, along with favorability values and confidence values associated with different vehicle speeds for the vehicle and their expected effects on the vehicle being able to travel through the first traffic light without having to come to a complete stop, based at least in part on optimizing the travelling time by measuring the favorability of possible passing through the first traffic light with a green signal at a particular point in time, utilizing the machine learning in combination with the one or more optimization objective function under feasibility constraints; and
fine-grain speed planning, via machine learning applied by the processor to the sensor data with respect to the first traffic light, the second traffic light, and additional consecutive traffic lights along the roadway in which the vehicle is travelling.
|