AI-based positioning monitoring process

Abstract

A system, method, and circuit are provided for monitoring the performance of artificial intelligence / machine learning models used for positioning assistance. A device is provided that includes memory and a processor coupled to the memory. The processor is configured to execute instructions stored in memory, receive monitoring KPIs derived from AI / ML data associated with the AI / ML model used for AI-assisted positioning, compare the monitoring KPIs to a baseline KPI, and perform one or more monitoring actions based on the comparison.

Claims

1. A device comprising memory and a processor coupled to the memory, wherein when the processor executes an instruction stored in the memory, the device: One or more measurement entities are made to measure a reference signal and generate measurement data for use when monitoring artificial intelligence / machine learning (AI / ML) models used for AI-assisted positioning. The performance of the AI / ML model is monitored based on the aforementioned measurement data. It is structured in such a way. device.

2. The device according to claim 1, wherein the reference signal is transmitted in a pre-configured periodic, aperiodic, or pseudo-periodic manner.

3. The device according to claim 2, wherein the processor is configured to cause the device to transmit the configuration of the reference signal to one or more reference signal entities, the configuration of which the reference signal is periodic, aperiodic, or pseudoperiodic.

4. The device according to claim 1, wherein the processor is configured to cause the device to transmit a physical layer request for the reference signal to one or more reference signal entities.

5. The device according to claim 1, wherein the measurement data is used by the AI / ML model to generate inference results.

6. The processor provides the device, The system is configured such that one or more measurement entities store measurement data collected in response to multiple reference signals and transmit the stored measurement data to the device. The device according to claim 1.

7. The processor provides the device, Trigger one or more reference signal entities to transmit multiple reference signals, Triggering one or more measurement entities to measure the plurality of reference signals, It is structured in such a way. The device according to claim 1.

8. The device according to any one of claims 1 to 7, wherein the measurement data includes reference signal received power (RSRP) or signal-to-interference noise ratio (SINR) of a reference signal, delay spread between a plurality of reference signals, or Doppler shift of a reference signal, channel impulse response (CIR) of a reference signal, power delay profile (PDP) of a reference signal, time of arrival (TOA) of a reference signal, and angle of arrival (AOA) of a reference signal.

9. When an instruction stored in memory is executed, The device receives a dataset containing artificial intelligence / machine learning (AI / ML) data, wherein each AI / ML data within the dataset is based on measurement data generated by one or more measurement entities in response to one or more reference signals. To generate key performance indicators (KPIs) based on the aforementioned AI / ML data, The device is made to provide the monitoring KPI to a device that functions as a monitoring comparison entity that monitors the AI / ML model used for AI-assisted positioning, A processor configured to perform operations that include [specific actions].

10. The aforementioned operation, To derive each intermediate data from each part of the aforementioned dataset, The monitoring KPI is generated based on the aforementioned intermediate data, including, The processor according to claim 9.

11. The processor according to claim 10, wherein the operation includes deriving the intermediate data by performing a statistical analysis of a subset of the AI / ML data.

12. The aforementioned operation, The device is made to monitor a scenario change notification signal indicating the time when the operating scenario changed, The intermediate data is derived based on any received scenario change notification signal, including, The processor according to claim 10.

13. The processor according to claim 10, wherein the measurement data includes reference signal received power (RSRP) or the signal-to-interference noise ratio (SINR) of the reference signal, delay spread between a plurality of reference signals, or the Doppler shift of the reference signal, the channel impulse response (CIR) of the reference signal, the power delay profile (PDP) of the reference signal, the time of arrival (TOA) of the reference signal, and the angle of arrival (AOA) of the reference signal.

14. The aforementioned operation, To cause the aforementioned device to receive a continuous dataset, When a threshold number of datasets are received, the monitoring KPI is generated using the received consecutive datasets, including, The processor according to any one of claims 9 to 13.

15. The processor according to any one of claims 9 to 13, wherein the measurement data includes inference data generated by the AI / ML model.

16. A device comprising memory and a processor coupled to the memory, wherein when the processor executes an instruction stored in the memory, the device: The system receives monitoring KPI values ​​derived from AI / ML data associated with the AI / ML model used for AI-assisted positioning. The aforementioned monitoring KPI is compared with the reference KPI. Based on the above comparison, one or more monitoring actions are performed. It is structured in such a way. device.

17. The device according to claim 16, wherein the standard KPI is derived from training data associated with the AI / ML model.

18. The device according to claim 16, wherein the monitoring KPI is derived from the inference data output by the AI / ML model.

19. The processor provides the device, As the aforementioned standard KPI, the ground truth value was determined. The monitoring KPI is compared with the ground truth value. It is structured in such a way. The device according to claim 16.

20. The device according to any one of claims 16 to 19, wherein the one or more monitoring actions include triggering a full or detailed retraining of the AI / ML model, selecting a different AI / ML model, deactivating AI-assisted positioning, or signaling a decision entity based on the comparison.

Images