Automatic braking of autonomous vehicles using machine learning based prediction of behavior of a traffic entity
An autonomous vehicle uses machine learning based models to predict hidden context attributes associated with traffic entities. The system uses the hidden context to predict behavior of people near a vehicle in a way that more closely resembles how human drivers would judge the behavior. The system determines an activation threshold value for a braking system of the autonomous vehicle based on the hidden context. The system modifies a world model based on the hidden context predicted by the machine learning based model. The autonomous vehicle is safely navigated, such that the vehicle stays at least a threshold distance away from traffic entities.
1. A method comprising:
receiving, by an autonomous vehicle, sensor data from sensors mounted on the autonomous vehicle;
identifying, one or more traffic entities based on the sensor data, the one or more traffic entities representing non-stationary objects in traffic in which the autonomous vehicle is driving;
for each of the one or more traffic entities displayed in an input image or a video:
providing the input image or video to a machine learning based model, wherein the machine learning based model is trained based on user responses received from human observers presented with images or videos showing traffic scenarios;
executing the machine learning based model to predict a measure of statistical distribution of user responses expected from human observers presented with the input image or video; and
determining an activation threshold for braking system of the autonomous vehicle based on the predicted measure of statistical distribution of user responses output by the machine learning based model;
predicting that the autonomous vehicle is likely to reach within the activation threshold of a particular traffic entity within a threshold time interval; and
responsive to determining that the autonomous vehicle is likely to reach within the activation threshold of the particular traffic entity within the threshold time interval, activating the braking system of the autonomous vehicle.
2. The method of claim 1 , wherein the user responses predict a hidden context representing a state of mind of a user represented by a traffic entity.
3. The method of claim 1 , wherein the user responses predict a hidden context representing a task that a user represented by a traffic entity is planning on accomplishing.
4. The method of claim 1 , wherein the user responses predict a hidden context representing a degree of awareness of the autonomous vehicle by a user represented by a traffic entity.
5. The method of claim 1 , wherein the user responses predict a hidden context representing a goal of a user represented by a traffic entity, wherein the user expects to achieve the goal within a threshold time interval.
6. The method of claim 1 , wherein the user responses predict a hidden context and wherein navigating the autonomous vehicle comprises:
generating signals for controlling the autonomous vehicle based on one or more motion parameters and the hidden context of each of the one or more traffic entities; and
sending the generated signals to controls of the autonomous vehicle.
7. The method of claim 1 , wherein the sensor data represents one or more images captured by a camera mounted on the autonomous vehicle.
8. The method of claim 1 , wherein the sensor data represents a lidar scan captured by a lidar mounted on the autonomous vehicle.
9. The method of claim 1 , wherein the machine learning based model predicts a kurtosis of a distribution of the user responses expected from the human observers presented with the input image or video.
10. The method of claim 1 , wherein the machine learning based model predicts a variance of a distribution of the user responses expected from the human observers presented with the input image or video.
11. A non-transitory computer readable storage medium storing instructions that when executed by a processor, cause the processor to perform steps comprising:
receiving, by an autonomous vehicle, sensor data from sensors mounted on the autonomous vehicle;
identifying, one or more traffic entities based on the sensor data, the one or more traffic entities representing non-stationary objects in traffic in which the autonomous vehicle is driving;
for each of the one or more traffic entities displayed in an input image or a video:
providing the input image or video to a machine learning based model, wherein the machine learning based model is trained based on user responses received from human observers presented with images or videos showing traffic scenarios;
executing the machine learning based model to predict a measure of statistical distribution of user responses expected from human observers presented with the input image or video; and
determining an activation threshold for braking system of the autonomous vehicle based on the predicted measure of statistical distribution of user responses output by the machine learning based model;
predicting that the autonomous vehicle is likely to reach within the activation threshold of a particular traffic entity within a threshold time interval; and
responsive to determining that the autonomous vehicle is likely to reach within the activation threshold of the particular traffic entity within the threshold time interval, activating the braking system of the autonomous vehicle.
12. The non-transitory computer readable storage medium of claim 11 , wherein the user responses predict a hidden context representing a state of mind of a user represented by a traffic entity.
13. The non-transitory computer readable storage medium of claim 11 , wherein the user responses predict a hidden context representing a task that a user represented by a traffic entity is planning on accomplishing.
14. The non-transitory computer readable storage medium of claim 11 , wherein the user responses predict a hidden context representing a degree of awareness of the autonomous vehicle by a user represented by a traffic entity.
15. The non-transitory computer readable storage medium of claim 11 , wherein the user responses predict a hidden context representing a goal of a user represented by a traffic entity, wherein the user expects to achieve the goal within a threshold time interval.
16. A computer system comprising:
a processor; and
a non-transitory computer readable storage medium storing instructions that when executed by the processor, cause the processor to perform steps comprising:
receiving, by an autonomous vehicle, sensor data from sensors mounted on the autonomous vehicle;
identifying, one or more traffic entities based on the sensor data, the one or more traffic entities representing non-stationary objects in traffic in which the autonomous vehicle is driving;
for each of the one or more traffic entities displayed in an input image or a video:
providing the input image or video to a machine learning based model, wherein the machine learning based model is trained based on user responses received from human observers presented with images or videos showing traffic scenarios;
executing the machine learning based model to predict a measure of statistical distribution of user responses expected from human observers presented with the input image or video; and
determining an activation threshold for braking system of the autonomous vehicle based on the predicted measure of statistical distribution of user responses output of the machine learning based model;
predicting that the autonomous vehicle is likely to reach within the activation threshold of a particular traffic entity within a threshold time interval; and
responsive to determining that the autonomous vehicle is likely to reach within the activation threshold of the particular traffic entity within the threshold time interval, activating the braking system of the autonomous vehicle.
17. The computer system of claim 16 , wherein the user responses predict a hidden context representing a state of mind of a user represented by a traffic entity.
18. The computer system of claim 16 , wherein the user responses predict a hidden context representing a task that a user represented by a traffic entity is planning on accomplishing.
19. The computer system of claim 16 , wherein the user responses predict a hidden context representing a degree of awareness of the autonomous vehicle by a user represented by a traffic entity.