Vehicle controller, method, and computer program for vehicle trajectory planning and control based on other vehicle behavior

Patent metadata

Field	Value
Publication number	US-11685405-B2
Application number	US-202117218280-A
Country	US
Kind code	B2
Filing date	Mar 31, 2021
Priority date	Apr 6, 2020
Publication date	Jun 27, 2023
Grant date	Jun 27, 2023

Abstract

A vehicle controller includes a processor configured to detect an object region including another vehicle near a vehicle from each of time series images obtained by a camera mounted on the vehicle; detect a predetermined action taken by the other vehicle, based on a trajectory of the other vehicle estimated from the object region of each image; identify the state of a signal light of the other vehicle, based on characteristics obtained from pixel values of the object region of each image; extract information indicating characteristics of an action of the other vehicle or the state of a signal light at the predetermined action taken by the other vehicle, based on the predetermined action detected in a tracking period and the state of the signal light related to the predetermined action; and predict behavior of the other vehicle, using the extracted information.

First claim

What is claimed is:

1. A vehicle controller for controlling travel of a vehicle, comprising:
a processor configured to: input time series images obtained by a camera mounted on the vehicle into a first classifier to detect an object region including another vehicle near the vehicle in each of the time series images;
track the other vehicle to estimate a trajectory of the other vehicle, based on the object region of each of the time series images;
detect a predetermined action taken by the other vehicle, based on the determined trajectory;
input characteristics obtained from pixel values of the object region detected in each of the time series images into a second classifier to identify the state of a signal light of the other vehicle, the second classifier having a recursive structure or performing a convolution operation in a temporal direction;
extract at least one piece of behavior-characteristic information indicating characteristics of an action of the other vehicle or the state of a signal light of the other vehicle at the predetermined action taken by the other vehicle, based on the predetermined action detected in a tracking period during which the other vehicle is tracked and the state of the signal light of the other vehicle identified in the tracking period and related to the predetermined action, wherein the at least one piece of behavior-characteristic information comprises a rate of nonperformance of lane changes that is a ratio of the number of times the other vehicle does not change lanes after blinking of a turn signal to the number of times the other vehicle makes a turn signal blink in the tracking period;
predict behavior of the other vehicle, using the at least one piece of behavior-characteristic information;
create a trajectory to be traveled on which the vehicle will travel, based on the predicted behavior of the other vehicle; and
control the vehicle so that the vehicle will travel along the trajectory to be traveled.

2. The vehicle controller according to claim 1 , wherein the processor determines, based on the trajectory of the other vehicle, a predicted trajectory along which the other vehicle is predicted to pass from the current time to a predetermined time ahead, generates a cost map in which costs of locations on the predicted trajectory are set higher than costs of other locations, corrects the cost map, using the at least one piece of behavior-characteristic information, to obtain a corrected cost map representing the predicted behavior of the other vehicle, and creates the trajectory to be traveled in accordance with the corrected cost map so that costs on the trajectory to be traveled will be the lowest.

3. The vehicle controller according to claim 2 , wherein the processor detects a lane change made by the other vehicle as the predetermined action taken by the other vehicle, and determines changing a turn signal of the other vehicle into a blinking state as changing into the state of the signal light related to the predetermined action.

4. The vehicle controller according to claim 3 , wherein in the case that the processor has inputted characteristics of the object region in the latest image obtained by the camera after the time series images into the second classifier and thereby recognized that a turn signal of the other vehicle has changed into a blinking state, the processor corrects the cost map so that costs of individual locations which the other vehicle assumed to change lanes to a direction indicated by the turn signal will pass are increased by a first value when the rate of nonperformance of lane changes is less than a first threshold and corrects the cost map so that the costs of the individual locations are increased by a second value smaller than the first value when the rate of nonperformance of lane changes is not less than the first threshold.

5. The vehicle controller according to claim 3 , wherein the processor extracts, as one of the at least one piece of behavior-characteristic information, a rate of lighting-up of turn signals that is a ratio of the number of times the other vehicle makes a turn signal blink before changing lanes to the number of times the other vehicle changes lanes in the tracking period.

6. The vehicle controller according to claim 5 , wherein in the case that the processor has inputted characteristics of the object region in the latest image obtained by the camera after the time series images into the second classifier and thereby recognized that turn signals of the other vehicle are off, the processor corrects the cost map so that costs of individual locations which are not on the predicted trajectory are decreased by a predetermined value when the rate of lighting-up of turn signals is not less than a second threshold and corrects the cost map so that costs of individual locations which the other vehicle assumed to change lanes will pass are equal to the costs of the locations on the predicted trajectory when the rate of lighting-up of turn signals is less than the second threshold.

7. The vehicle controller according to claim 3 , wherein the processor extracts, as one of the at least one piece of behavior-characteristic information, a rate of occurrence of lane changes that is the number of lane changes made by the other vehicle per unit time.

8. The vehicle controller according to claim 7 , wherein the processor corrects the cost map so that costs of individual locations which are not on the predicted trajectory are decreased by a predetermined value when the rate of occurrence of lane changes is not greater than a third threshold and corrects the cost map so that costs of individual locations which the other vehicle assumed to change lanes will pass are equal to the costs of the locations on the predicted trajectory when the rate of occurrence of lane changes is greater than the third threshold.

9. The vehicle controller according to claim 3 , wherein the processor extracts, as one of the at least one piece of behavior-characteristic information, an average of time lags each of which is a period from when the other vehicle starts blinking a turn signal before a lane change in the tracking period until the other vehicle changes lanes.

10. The vehicle controller according to claim 9 , wherein upon input of the characteristics obtained from pixel values of the object region, the second classifier outputs a confidence score indicating how likely a turn signal that is a signal light of the other vehicle to be blinking, and
the processor determines that the turn signal is blinking, when the confidence score is not less than a confidence-score threshold, and decreases the confidence-score threshold by a predetermined value when the average of the time lags is less than a predetermined threshold.

11. The vehicle controller according to claim 1 , wherein the vehicle further includes a second camera configured to capture a second capturing region to generate a second image, the second capturing region being different from a capturing region of the camera,
the processor inputs the second image into the first classifier to detect an object region including the other vehicle in the second image, and compares the object region in the second image with the object region in one of the time series images to track the other vehicle.

12. The vehicle controller according to claim 1 , wherein the vehicle further includes a range sensor configured to generate a ranging signal every predetermined period, the ranging signal including, for each angle in a predetermined range of angles around the vehicle, a measured value of a distance to an object located in the direction of the angle, and
the processor determines a range of

Assignees

• Toyota Motor Co Ltd
• Denso Corp

Inventors

• Hashimoto Daisuke
• Kuroda Ryusuke
• Sugiura Tetsuro
• Kamada Hitoshi

Classifications

• B60W2420/403

  Image sensing, e.g. optical camera · CPC title

• G06V10/25

  Determination of region of interest [ROI] or a volume of interest [VOI] · CPC title

• B60W2554/4045

  Intention, e.g. lane change or imminent movement · CPC title

• B60W30/08

  Active safety systems} predicting or avoiding probable or impending collision {or attempting to minimise its consequences · CPC title

• B60W60/00272 Primary

  relying on extrapolation of current movement · CPC title