System, method, and computer program product for data-driven optimization of onboard data collection

1. A computer-implemented method for onboard data collection, comprising: identifying, by at least one processor, a condition in a roadway associated with an operation of one or more autonomous vehicles (AVs) that may be further optimized to improve performance of the one or more AVs, identifying the condition in the roadway further includes:

mining, by at least one processor, a plurality of roadway images from one or more data sources for factors relating to a reoccurring situation, as the condition, in the roadway; and

associating, by the at least one processor, the plurality of roadway images with one or more other roadways to traverse based on location and time; and

generating, by the at least one processor, one or more condition capture instructions to communicate to the one or more AVs, wherein the one or more condition capture instructions comprise one or more parameters of a storage request in one or more roadways predicted to exhibit the condition in the roadway,

wherein, when at least one condition capture instruction in the one or more condition capture instructions are received by an AV computing system in at least one of the one or more AVs, the at least one condition capture instruction is configured to cause the at least one of the one or more AVs to collect condition information in the roadway at a time when the condition is predicted to be present in the roadway;

controlling at least one of the one or more AVs based on the one or more condition capture instructions.

2. The computer-implemented method of claim 1 , wherein generating the one or more condition capture instructions comprises:

checking the confidence interval of a specific metric associated with the condition in the roadway to determine whether each of the one or more roadways is in a specific target range;

determining a confidence interval of a specific metric associated with the condition in the roadway is in a specific target range for at least one roadway of the one or more roadways; and

in response to determining the confidence interval is within a specific target range for the at least one roadway,

transmitting, to the at least one of the one or more AVs, an instruction in the one or more condition capture instructions to collect condition information in the at least one roadway.

3. The computer-implemented method of claim 1 , wherein the condition in the roadway is determined to be present by determining target information based on one or more objects identified in information previously collected which also include the condition in the roadway, and the target information includes parameters for at least one of a place, a location, or a time, for controlling logging information when the condition in the roadway is predicted to be present.

4. The computer-implemented method of claim 1 , further comprising:

determining a time or a location when a first AV of the one or more AVs may traverse a roadway having an observable condition;

determining a time or a location when a second AV of the one or more AVs may traverse a roadway having an observable condition related to at least one of a place, a location, or a time, that should be logged;

matching roadway information from one or more sources of information that are related to at least one of the condition in the roadway; and

sending an instruction to the one or more AVs to cause the first AV and the second AV to log information of the roadway in an environment surrounding each respective AV for a specified location at the time when the condition is predicted to be present in the roadway.

5. The computer-implemented method of claim 4 , further comprising:

determining that a threshold number of AVs have previously collected information about the specified place, the location, or the time; and

sending an instruction to a fleet of AVs to cause each of the one or more AVs of the fleet of AVs to stop logging information for the specified place, the location, or the time.

6. The computer-implemented method of claim 1 , wherein the condition in the roadway includes a naturally occurring situation in and around a roadway that may affect one or more operations of the at least one of the one or more AVs, execution, movement, or action, and

wherein at least one signal, factor, or object serves as a proxy for the naturally occurring condition.

7. The computer-implemented method of claim 1 , wherein the onboard data collection is optimized by performing at least one of: obtaining a specified number of images, obtaining a number of images to perform a specific process, or obtaining a number of images to update an inference engine.

8. The computer-implemented method of claim 1 , wherein the onboard data collection is optimized by eliminating or decreasing processing for logging, review, or analysis of objects,

wherein logging comprises eliminating extra logging by logging based on AV mining of: images that improve perception of autonomous systems or contribute to solving problems associated with navigating in the roadway.

9. A system, comprising: a memory; and at least one processor coupled to the memory and configured to:

identify a condition in a roadway associated with an operation of one or more autonomous vehicles (AVs) that may be further optimized to improve performance of the one or more AVs, identify the condition in the roadway further includes:

mine a plurality of roadway images from one or more data sources for factors relating to or matching a reoccurring situation, as the condition, in the roadway; and

associate the plurality of roadway images with one or more other roadways to traverse based on location and time; and

generate one or more condition capture instructions to communicate to the one or more AVs, one or more parameters of a storage request in one or more roadways predicted to exhibit the condition in the roadway,

wherein, when at least one condition capture instruction in the one or more condition capture instructions are received by an AV computing system in at least one of the one or more AVs, the at least one condition capture instruction is configured to cause the at least one AV of the one or more AVs to collect condition information in the roadway at a time when the condition is predicted to be present in the roadway,

control at least one of the one or more AVs based on the one or more condition capture instructions.

10. The system of claim 9 , wherein the at least one processor is further configured to:

check the confidence interval of a specific metric associated with the condition in the roadway to determine whether each of the one or more roadways is in a specific target range;

determine a confidence interval of a specific metric associated with the condition in the roadway is in a specific target range for at least one roadway of the one or more roadways; and

in response to determining the confidence interval is within a specific target range for the at least one roadway,

transmit, to the at least one AV of the one or more AVs, an instruction in the one or more condition capture instructions to collect condition information in the at least one roadway.

11. The system of claim 9 , wherein the condition in the roadway is determined to be present by determining target information based on one or more objects identified in information previously collected which also include the condition in the roadway, and the target information includes parameters for at least one of a place, a location, or a time, for controlling logging information when the condition in the roadway is predicted to be present.

12. The system of claim 9 , wherein the at least one processor is further configured to:

determine a time or a location when a first AV of the one or more A Vs may traverse a roadway having an observable condition;

determine a time or a location when a second AV of the one or more AVs may traverse a roadway having an observable condition related to at least one of a place, a location, or a time, that should be logged;

match roadway information from one or more sources of information that are related to at least one of the condition in the roadway; and

send an instruction to the one or more AVs to cause the first AV and the second AV to log information of the roadway in an environment surrounding each respective AV for a specified location at the time when the condition is predicted to be present in the roadway.

13. The system of claim 12 , wherein the at least one processor is configured to:

determine that a threshold number of AVs have previously collected information about the specified place, location, or time; and

send an instruction to a fleet of AVs to cause each of the one or more AVs of the fleet of AVs to stop logging information for the specified place, location, or time.

14. The system of claim 9 , wherein the condition in the roadway includes a naturally occurring situation in and around a roadway that may affect one or more operations of the at least one AV of the one or more AVs, execution, movement, or action, and

wherein at least one signal, factor, or object serves as a proxy for the naturally occurring condition.

15. The system of claim 9 , wherein the processor is further configured to perform at least one of: obtaining a specified number of images, obtaining a number of images to perform a specific process, or obtaining a number of images to update an inference engine.

16. The system of claim 9 , wherein the processor is further configured to eliminate or decrease processing for logging, review, or analysis of objects,

wherein logging comprises eliminating extra logging by logging based only on AV mining, comprising: collecting images that improve perception of autonomous systems or contribute to solving problems associated with navigating in the roadway.

17. A non-transitory computer-readable medium having instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to:

identify a condition in a roadway associated with an operation of one or more autonomous vehicles (AVs) that may be further optimized to improve performance of the one or more AVs, identify the condition in the roadway further includes:

mine a plurality of roadway images from one or more data sources for factors relating to (or matching) a reoccurring situation in the roadway; and

associate the plurality of roadway images with one or more other roadways to traverse based on location and time; and

generate one or more condition capture instructions to communicate to the one or more AVs, one or more parameters of a storage request in one or more roadways predicted to exhibit the condition in the roadway;

wherein, when at least one condition capture instruction in the one or more condition capture instructions are received by an AV computing system in at least one of the one or more AVs, the at least one condition capture instruction is configured to cause the at least one AV of the one or more AVs to collect condition information in the roadway at a time when the condition is predicted to be present in the roadway,

control at least one of the one or more AVs based on the one or more condition capture instructions.

18. The non-transitory computer-readable medium of claim 17 , having further instructions stored thereon that, when executed by at least one computing device, cause the at least one computing device to:

check the confidence interval of a specific metric associated with the condition in the roadway to determine whether each of the one or more roadways is in a specific target range;

determine a confidence interval of a specific metric associated with the condition in the roadway is in a specific target range for at least one roadway of the one or more roadways; and

in response to determining the confidence interval is within a specific target range for the at least one roadway,

transmit, to the AV, an instruction in the one or more condition capture instructions to collect condition information in the at least one roadway.