Abstract
A user equipment (UE) in a non-standalone (NSA) 5G network uses a deep learning model to predict whether a better downlink (DL) throughput can be obtained by enabling 5G new radio (NR) capability to optimize its DL data throughput performance. 4G/5G measurable key performance indicators (KPIs) are recorded by the UE as training features during daily usage. The UE also records the perceived DL throughput for both the LTE-only mode and the 5G NR dual connectivity mode as labels. The deep learning model is trained based on the recorded features and labels using, for example, supervised learning. The UE implements the trained deep learning model to predict whether a better DL throughput can be obtained by enabling 5G NR dual connectivity based on the features, such as the measurable 4G/5G KPIs, at the current location.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Kuo, Gordon, "Optimizing Data Throughput Performance in 5G Non-Standalone Networks Using Deep Learning", Technical Disclosure Commons, (March 15, 2021)
https://www.tdcommons.org/dpubs_series/4151