JP-WO2025033472-A5 - Communication methods, user devices, mobile communication systems, programs, and chipsets

Dates

Publication Date

20260511

Application Date

20240807

Description

This disclosure relates to communication methods, user equipment, mobile communication systems, programs, and chipsets used in mobile communication systems. The configuration information may be information for configuring the application of inference processing to the prediction of the CSI on the downlink. For example, the configuration information may include a function ID indicating CSI prediction, or a model ID of the AI/ML model for CSI prediction. In step S102, UE100 applies inference processing to predict CSI only when inference processing is required (application timing). That is, UE100 applies the AI/ML model for CSI prediction to predict CSI at the application timing. This reduces the processing load of the inference processing for CSI prediction in UE100. Note that at non-application timings, UE100 performs CSI measurement in the same manner as before. In this case, gNB200 must transmit CSI-RS (full CSI-RS) in the same manner as before. On the other hand, gNB200 can simplify (e.g., transmit punctured CSI-RS) or stop transmitting CSI-RS to UE100 at the timing when inference processing is required (application timing) because UE100 applies the AI/ML model, thereby saving radio resources. gNB200 may also set the timing for stopping CSI-RS transmission to UE100 as the timing when inference processing is required (application timing). In the first operation pattern , the configuration information includes the discontinuous operation settings described above. The configuration information may also include identification information for the AI/ML model. The configuration information may also include information indicating the frequency resource (e.g., frequency band, BWP) to which the model inference will be applied. The configuration information may also include information indicating the space (e.g., tracking area, arbitrary geographical location) to which the model inference will be applied. The configuration information may also include settings related to the inference data (input data to the model). These settings may include information indicating how many past CSI measurement results will be used as input to the AI/ML model (e.g., a minimum of 10).