Computer-assisted or autonomous driving traffic sign recognition method and apparatus

A method for operating an adversarial resilience subsystem (ARS), the method comprising:

• obtaining, by the ARS, a classification of a traffic sign, a confidence level associated with the classification, and a location of the traffic sign during operation of a CA/AD vehicle, wherein the classification of the traffic sign is generated based at least in part on computer vision;
• querying, by the ARS, a remote sign locator service for a reference description of the traffic sign at the location based on the obtained classification of the traffic sign;
• obtaining, by the ARS during operation of the CA/AD vehicle, the reference description of the traffic sign, wherein the reference description is based on a plurality of traffic sign classifications at or near the location;
• detecting, by the ARS, anomalies between the classification of the traffic sign and the reference description of the traffic sign;
• and
• determining, by the ARS, whether the classification is correct or incorrect based at least in part on an amount of anomalies detected and the confidence level associated with the classification.

One or more non-transitory computer-readable media (NTCRM) comprising instructions of an adversarial resilience subsystem (ARS), wherein execution of the instructions by one or more processors of a computing device is to cause the computing device to:

• obtain, during operation of a CA/AD vehicle, a classification of a traffic sign, a confidence level associated with the classification, and a location of the traffic sign, wherein the classification of the traffic sign is generated based at least in part on computer vision;
• query a remote sign locator service for a reference description of the traffic sign at the location based on the obtained classification of the traffic sign;
• obtain the reference description of the traffic sign in response to the query, wherein the reference description is based on a plurality of traffic sign classifications at or near the location;
• detect anomalies based on a comparison of the reference description with the classification of the traffic sign and the location of the traffic sign;
• and
• determine whether the classification is correct or incorrect based at least in part on an amount of anomalies detected and the confidence level associated with the classification.