Antenna monitoring and selection

1. A method, comprising:

obtaining antenna performance data for a set of antennas of an autonomous vehicle;

determining a real-time connectivity score for each antenna of the set of antennas based on the antenna performance data, wherein determining the real-time connectivity score for each antenna of the set of antennas based on the antenna performance data comprises:

obtaining weights for at least a portion of the antenna performance data corresponding to the respective antenna,

determining a weighted sum of the at least a portion of the antenna performance data corresponding to the respective antenna based on values of the at least a portion of the antenna performance data and the weights, and

determining the real-time connectivity score for the respective antenna based on the weighted sum of the at least a portion of the antenna performance data;

obtaining location data and environment data for the autonomous vehicle;

inputting the location data and the environment data into a machine learning model system;

receiving a predicted connectivity score for each antenna of the set of antennas from the machine learning model system; and

determining at least one antenna of the set of antennas to use for communication based on the predicted connectivity score and the real-time connectivity score for each antenna of the set of antennas.

2. The method of claim 1 , wherein the antenna performance data comprising at least one of antenna signal strength, uplink or downlink bandwidth, GNSS signal, latency, signal-to-noise ratio, or cell tower identification.

3. The method of claim 1 , wherein the location data comprises at least one of a location of the autonomous vehicle based on GNSS signals or a position on a map determined by a localization system.

4. The method of claim 1 , wherein the environment data comprises at least one of weather data obtained from a remote system or sensed onboard the autonomous vehicle, traffic data obtained from the remote system or sensed onboard the autonomous vehicle, or local contextual data obtained from the remote system.

5. The method of claim 4 , wherein the local contextual data includes at least a connectivity user density.

6. The method of claim 1 , wherein inputting the location data and the environment data to the machine learning model system comprises:

generating a feature vector based on the location data and the environment data; and

inputting the feature vector to the machine learning model system, wherein the machine learning model system comprises a scoring model trained to predict scores based on feature vectors.

7. The method of claim 6 , wherein the scoring model is trained using a training dataset, wherein the training dataset comprises historical connectivity scores, historical location data, and historical environment data based on fleet data from a plurality of autonomous vehicles, wherein training the scoring model using the training dataset comprises training the scoring model to predict the historical connectivity scores based on feature vectors generated based on the historical location data and historical environment data.

8. A method, comprising:

obtaining antenna performance data for a set of antennas of an autonomous vehicle;

determining a real-time connectivity score for each antenna of the set of antennas based on the antenna performance data;

obtaining location data and environment data for the autonomous vehicle;

inputting the location data and the environment data into a machine learning model system;

receiving a predicted connectivity score for each antenna of the set of antennas from the machine learning model system; and

determining at least one antenna of the set of antennas to use for communication based on the predicted connectivity score and the real-time connectivity score for each antenna of the set of antennas, wherein determining the at least one antenna of the set of antennas to use for communication comprises:

determining a first predicted connectivity score associated with a first antenna of the set of antennas and a first real-time connectivity score associated with the first antenna satisfy a connectivity threshold; and

in response to the connectivity threshold being satisfied for the first antenna, selecting a second antenna to use for communication that does not satisfy the connectivity threshold.

9. The method of claim 8 , wherein the connectivity threshold is satisfied when the first real-time connectivity score is less than the first predicted connectivity score, when the first real-time connectivity score is less than a lower range from the first predicted connectivity score, or when the first real-time connectivity score is less than a minimum connectivity score.

10. The method of claim 8 , further comprising:

in response to the connectivity threshold being satisfied for the first antenna, iterating a count of a counter associated with the first antenna;

determining the count of the counter is greater than a counter threshold; and

in response to determining the count of the counter is greater than the counter threshold, setting a diagnostic trouble code for the first antenna or determining to conduct a minimum risk maneuver.

11. The method of claim 8 , further comprising:

in response to the connectivity threshold being satisfied for the first antenna, transmitting autonomous vehicle data to a remote system, wherein the autonomous vehicle data comprises the antenna performance data or the real-time connectivity score in association with the location data and the environment data.

12. A system, comprising:

at least one processor, and

at least one non-transitory storage media storing computer-executable instructions that, when executed by the at least one processor, cause the at least one processor to:

obtain antenna performance data for a set of antennas of an autonomous vehicle;

determine a real-time connectivity score for each antenna of the set of antennas based on the antenna performance data, wherein to determine the real-time connectivity score for each antenna of the set of antennas based on the antenna performance data, the computer-executable instructions further cause the at least one processor to:

obtain weights for at least a portion of the antenna performance data corresponding to the respective antenna,

determine a weighted sum of the at least a portion of the antenna performance data corresponding to the respective antenna based on values of the at least a portion of the antenna performance data and the weights, and

determine the real-time connectivity score for the respective antenna based on the weighted sum of the at least a portion of the antenna performance data;

obtain location data and environment data for the autonomous vehicle;

input the location data and the environment data to a machine learning model system;

receive a predicted connectivity score for each antenna of the set of antennas from the machine learning model system; and

determine at least one antenna of the set of antennas to use for communication based on the predicted connectivity score and the real-time connectivity score for each antenna of the set of antennas.

13. The system of claim 12 , wherein the antenna performance data comprising at least one of antenna signal strength, uplink or downlink bandwidth, GNSS signal, latency, signal-to-noise ratio, or cell tower identification.

14. The system of claim 12 , wherein the location data comprises at least one of a location of the autonomous vehicle based on GNSS signals or a position on a map determined by a localization system.

15. The system of claim 12 , wherein the environment data comprises at least one of weather data obtained from a remote system or sensed onboard the autonomous vehicle, traffic data obtained from the remote system or sensed onboard the autonomous vehicle, or local contextual data obtained from the remote system.

16. The system of claim 12 , wherein to input the location data and the environment data to the machine learning model system, the computer-executable instructions further cause the at least one processor to:

generate a feature vector based on the location data and the environment data; and

input the feature vector to the machine learning model system, wherein the machine learning model system comprises a scoring model trained to predict scores based on feature vectors.

17. The system of claim 16 , wherein the scoring model is trained using a training dataset, wherein the training dataset comprises historical connectivity scores, historical location data, and historical environment data based on fleet data from a plurality of autonomous vehicles, wherein training the scoring model using the training dataset comprises training the scoring model to predict the historical connectivity scores based on feature vectors generated based on the historical location data and historical environment data.