Method for machine learning model monitoring management

Abstract

A method at an apparatus (e.g. a UE or network node) for inter-frequency handover (IFHO) measurements comprising measuring a first characteristic of a first layer of a wireless communication system; predicting, based on an inference in a first model (e.g. a machine learning (ML) model), a success of a transfer from the first layer to a second layer (e.g. transfer between macrocell and microcell). Based on the inference, and on a configuration parameter that indicates a second characteristic is to be measured, measuring the second characteristic of the wireless communication system, wherein the second characteristic is a characteristic of the second layer and wherein measuring the second characteristic is based on the inference of transfer success from the first layer to the second layer being positive. Finally, updating the first model based on the measuring of the second characteristic being performed. The first and second characteristics may be reference signal receive power (RSRP). The configuration parameter may be a probability parameter or model monitoring window. The configuration parameter may be based on an amount of detected drift or degradation.

Images