Method and system for detecting faulty devices

1. A method for detecting faulty devices, the method comprising the steps of:

gathering test data in near field with respect to a device under test,

extrapolating the test data to far field conditions with the aid of at least one machine learning technique, and

evaluating a far field performance of the device under test on the basis of the far field conditions,

wherein a number of measurement antennas being employed for gathering the test data is reduced by a factor of at least 10 in comparison to a reconstructed far field measurement having been extrapolated on the basis of a near field measurement with the aid of a nonmachine learning technique.

2. The method according to claim 1 , wherein the test data is sparse test data.

3. The method according to claim 2 ,

wherein the sparse test data comprises less information than reconstruction data being necessary for reconstructing a far field behavior of the device under test with the aid of a non-machine learning technique, or

wherein the sparse test data is recorded during production of the device under test, or with limited resources.

4. The method according to claim 3 ,

wherein the non-machine learning technique comprises at least one of a spherical wave expansion, a plane wave expansion, or a multiple expansion.

5. The method according to claim 1 , wherein gathering the test data is done in an over-the-air manner.

6. The method according to claim 1 ,

wherein gathering the test data employs less than 100 measurement antennas, less than 50 measurement antennas, less than 20 measurement antennas, or less than 10 measurement antennas.

7. The method according to claim 1 ,

wherein the number of measurement antennas being employed for gathering the test data is reduced by a factor of 10 in comparison to a far field measurement with respect to the device under test.

8. The method according to claim 1 ,

wherein the method further comprises the step of recalibrating the device under test on the basis of the far field conditions selected from a calibration error in magnitude, a calibration error in phase, or a calibration error in magnitude and phase.

9. The method according to claim 1 ,

wherein the method further comprises the step of training the at least one machine learning technique on the basis of at least one of simulated data, approximated data, predefined data, or real data.

10. The method according to claim 1 ,

wherein the method further comprises the step of training the at least one machine learning technique on the basis of real data measured at a production line of devices under test.

11. The method according to claim 10 ,

wherein the step of training the at least one machine learning technique on the basis of real data measured at a production line of devices under test comprises extrapolating near field data at production site to far field data.

12. The method according to claim 10 ,

wherein the step of training the at least one machine learning technique on the basis of real data measured at a production line of devices under test comprises:

getting near field data measured at production site in combination with quality data with respect to the device under test, and

training the at least one machine learning technique on the basis of at least one of the near field data or the quality data.

13. The method according to claim 10 ,

wherein the step of training the at least one machine learning technique on the basis of real data measured at a production line of devices under test comprises:

extrapolating near field data at production site to far field data,

getting the near field data measured at the production site in combination with quality data with respect to the device under test, and

training the at least one machine learning technique on the basis of at least one of the near field data, the far field data, or the quality data.

14. A system for detecting faulty devices, the system comprising:

a measurement equipment configured to gather test data in near field with respect to a device under test,

an extrapolator configured to extrapolate the test data to far field conditions with the aid of at least one machine learning technique, and

an evaluator configured to evaluate a far field performance of the device under test on the basis of the far field conditions,

wherein the measurement equipment comprises a number of measurement antennas being reduced by a factor of at least 10 in comparison to a reconstructed far field measurement having been extrapolated on the basis of a near field measurement with the aid of a nonmachine learning technique.

15. The system according to claim 14 ,

wherein the test data is sparse test data, or

wherein the test data or the sparse test data comprises less information than reconstruction data being necessary for reconstructing a far field behavior of the device under test with the aid of a non-machine learning technique, or

wherein the test data or the sparse test data is recorded in a short amount of time, preferably during production of the device under test, and/or with limited resources, or

wherein the non-machine learning technique comprises at least one of a spherical wave expansion, a plane wave expansion, or a multiple expansion.

16. The system according to claim 14 ,

wherein the measurement equipment gathers the test data in an over-the-air manner, or

wherein the measurement equipment comprises less than 100 measurement antennas, preferably less than 50 measurement antennas, more preferably less than 20 measurement antennas, most preferably less than 10 measurement antennas, or

wherein the measurement equipment comprises the number of measurement antennas being reduced by a factor of 10 in comparison to a far field measurement with respect to the device under test.

17. The system according to claim 14 ,

wherein the evaluator is further configured to recalibrate the device under test on the basis of the far field conditions selected from a calibration error in magnitude, a calibration error in phase, or a calibration error in magnitude and phase.

18. The system according to claim 14 ,

wherein the system further comprises a training processor configured to train the at least one machine learning technique on the basis of at least one of simulated data, approximated data, predefined data, or real data.

19. The system according to claim 14 ,

wherein the system further comprises a training processor configured to train the at least one machine learning technique on the basis of real data measured at a production line of devices under test.

20. The system according to claim 19 ,

wherein the training processor is further configured to receive near field data measured at production site in combination with quality data with respect to the device under test from the measurement equipment, and

wherein the training processor is further configured to train the at least one machine learning technique on the basis of at least one of the near field data or the quality data.