Machine learning based handover parameter optimization
Disclosed is a method comprising obtaining a plurality of handover parameter values, using a first machine learning model to select a subset of handover parameter values from the plurality of handover parameter values, obtaining historical information of a plurality of terminal devices, determining a first set of optimal handover parameter values for the plurality of terminal devices from the subset of handover parameter values, tagging the first set of optimal handover parameter values with the historical information of the plurality of terminal devices to obtain a labelled dataset, and training a second machine learning model with the labelled dataset, wherein the trained second machine learning model is capable of predicting a second set of optimal handover parameter values for a first terminal device based on historical information of the first terminal device.
1. An apparatus, comprising:
at least one processor; and
at least one memory including computer program code, wherein the at least one memory and the computer program code are configured, with the at least one processor, to cause the apparatus to:
obtain a plurality of handover parameter values;
use a first machine learning model to select a subset of handover parameter values from the plurality of handover parameter values;
obtain historical information of a plurality of first terminal devices;
determine a first set of optimal handover parameter values for the plurality of first terminal devices from the subset of handover parameter values;
tag the first set of optimal handover parameter values with the historical information of the plurality of first terminal devices to obtain a labelled dataset; and
train a second machine learning model with the labelled dataset, wherein the trained second machine learning model is capable of predicting a second set of optimal handover parameter values for a second terminal device based on historical information of the second terminal device.
2. The apparatus according to claim 1 , wherein the at least one memory and computer program code are further configured, with the at least one processor, to cause the apparatus to provide the trained second machine learning model to one or more base stations.
3. The apparatus according to claim 1 , wherein the handover parameter values comprise values representing at least one of hysteresis margin and time-to-trigger.
4. The apparatus according to claim 1 , wherein the historical information comprises a cell ID, a beam ID, or a reference signal received power value.
5. The apparatus according to claim 1 , wherein the first set of optimal handover parameter values is determined by testing the subset of handover parameter values with at least a subset of the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the at least subset of the plurality of first terminal devices.
6. The apparatus according to claim 1 , wherein the first set of optimal handover parameter values is determined by simulating the subset of handover parameter values with the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the plurality of first terminal devices based on the simulation.
7. The apparatus according to claim 1 , wherein the apparatus comprises a device in a self-organizing network or a radio intelligent controller.
8. A system comprising one or more apparatuses, wherein the system is configured to:
obtain a plurality of handover parameter values;
use a first machine learning model to select a subset of handover parameter values from the plurality of handover parameter values;
obtain historical information of a plurality of first terminal devices;
determine a first set of optimal handover parameter values for the plurality of first terminal devices from the subset of handover parameter values;
tag the first set of optimal handover parameter values with the historical information of the plurality of first terminal devices to obtain a labelled dataset; and
train a second machine learning model with the labelled dataset, wherein the trained second machine learning model is capable of predicting a second set of optimal handover parameter values for a second terminal device based on historical information of the second terminal device.
9. The system according to claim 8 , further comprising providing the trained second machine learning model to one or more base stations.
10. The system according to claim 8 , wherein the handover parameter values comprise values representing at least one of hysteresis margin and time-to-trigger.
11. The system according to claim 8 , wherein the historical information comprises a cell ID, a beam ID, or a reference signal received power value.
12. The system according to claim 8 , wherein the first set of optimal handover parameter values is determined by testing the subset of handover parameter values with at least a subset of the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the at least subset of the plurality of first terminal devices.
13. The system according to claim 8 , wherein one of the apparatuses comprises a device in a self-organizing network or a radio intelligent controller.
14. The system according to claim 8 , wherein the first set of optimal handover parameter values is determined by simulating the subset of handover parameter values with the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the plurality of first terminal devices based on the simulation.
15. A method, comprising:
obtaining a plurality of handover parameter values;
using a first machine learning model to select a subset of handover parameter values from the plurality of handover parameter values;
obtaining historical information of a plurality of first terminal devices;
determining a first set of optimal handover parameter values for the plurality of first terminal devices from the subset of handover parameter values;
tagging the first set of optimal handover parameter values with the historical information of the plurality of first terminal devices to obtain a labelled dataset; and
training a second machine learning model with the labelled dataset, wherein the trained second machine learning model is capable of predicting a second set of optimal handover parameter values for a second terminal device based on historical information of the second terminal device.
16. The method according to claim 15 , further comprising providing the trained second machine learning model to one or more base stations.
17. The method according to claim 15 , wherein the handover parameter values comprise values representing at least one of hysteresis margin and time-to-trigger.
18. The method according to claim 15 , wherein the historical information comprises a cell ID, a beam ID, or a reference signal received power value.
19. The method according to claim 15 , wherein the first set of optimal handover parameter values is determined by testing the subset of handover parameter values with at least a subset of the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the at least subset of the plurality of first terminal devices.
20. The method according to claim 15 , wherein the first set of optimal handover parameter values is determined by simulating the subset of handover parameter values with the plurality of first terminal devices and selecting a set of handover parameter values that maximize a performance indicator of the plurality of first terminal devices based on the simulation.