Detecting and responding to malfunctioning traffic signals for autonomous vehicles
Aspects of the disclosure relate to detecting and responding to malfunctioning traffic signals for a vehicle having an autonomous driving mode. For instance, information identifying a detected state of a traffic signal for an intersection. An anomaly for the traffic signal may be detected based on the detected state and prestored information about expected states of the traffic signal. The vehicle may be controlled in the autonomous driving mode based on the detected anomaly.
1. A method of controlling a vehicle traveling along a route to a destination in an autonomous driving mode, the method comprising:
receiving, by one or more processors, information identifying a detected state of a traffic signal for an intersection;
detecting, by the one or more processors, an anomaly for the traffic signal based on the detected state and prestored information about expected states of the traffic signal;
identifying, by the one or more processors, a classification of the detected anomaly from one of a plurality of different classifications of anomaly types and a probability of the classification; and
controlling, by the one or more processors, the vehicle to make forward progress towards the destination in the autonomous driving mode based on the classification of the detected anomaly.
2. The method of claim 1 , wherein the information further identifies a detected location of the traffic signal and wherein detecting the anomaly is further based on the detected location.
3. The method of claim 1 , further comprising, receiving information about behavior of another object in an environment of the vehicle, and wherein detecting the anomaly is further based on the behavior.
4. The method of claim 1 , wherein detecting the anomaly is further based on whether the vehicle is within a threshold distance of the traffic signal where a state of the traffic signal should be detected.
5. The method of claim 4 , wherein detecting the anomaly is further based on a failure to detect the state of the traffic signal when the vehicle is within the threshold distance.
6. The method of claim 5 , wherein detecting the anomaly is further based on whether there are occlusions between the vehicle and the traffic signal.
7. The method of claim 5 , wherein the threshold distance is closer in distance or time to the traffic signal than a perceptive range of a system of the vehicle configured to detect a state of the traffic signal.
8. The method of claim 1 , wherein detecting the anomaly is further based on whether the vehicle has observed a same state of the traffic signal for more than a threshold period of time.
9. The method of claim 8 , further comprising, selecting the threshold period of time based on the detected state.
10. The method of claim 1 , wherein detecting the anomaly is further based on observed behaviors of other vehicles.
11. The method of claim 1 , further comprising:
receiving information about behavior of another object in an environment of the vehicle, wherein the detected anomaly is further classified based on the behavior, and wherein the controlling is further based on the further classification.
12. The method of claim 1 , further comprising classifying the anomaly based on whether the vehicle has observed a same state of the traffic signal for more than a threshold period of time, wherein the controlling is further based on the classification.
13. The method of claim 1 , wherein controlling the vehicle further includes accessing a mapping of anomaly classifications to vehicle responses and using the classification to determine a response from the mapping.
14. The method of claim 1 , wherein controlling the vehicle includes stopping at an intersection controlled by the traffic signal.
15. The method of claim 1 , wherein controlling the vehicle includes proceeding through an intersection controlled by the traffic signal without stopping.
16. The method of claim 1 , further comprising, observing behavior of cross-traffic, and wherein controlling the vehicle is further based on the observed behavior.
17. The method of claim 1 , wherein controlling the vehicle includes requesting assistance from a remote assistance operator before proceeding.
18. The method of claim 1 , wherein controlling the vehicle is further based on whether there is a pedestrian directing traffic at an intersection controlled by the traffic signal.
19. The method of claim 1 , wherein identifying the classification is based on a look up table which cross-references anomalies and ones of plurality of different types of anomalies.
20. The method of claim 1 , wherein the look up table further cross-references the anomalies and the ones of plurality of different classifications of anomaly types with responses, and wherein controlling the vehicle is further based on one of the responses.
21. The method of claim 1 , wherein classifying the anomaly includes using a machine learning model to determine the classification and the probability.