KPI Prediction Horizon Feedback for Wireless Link Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing approaches for link adaptation in wireless networks face challenges due to variance in time horizons for user equipment (UE) predictions, leading to inefficiencies such as trial and error in re-training and re-configuring machine learning models, and hindering resource allocation by radio base stations.
Innovation Solution
A method for adaptive prediction of time horizon for key performance indicators (KPIs) involving a feedback loop between network nodes and UEs, using machine learning models to transmit and update predicted time horizons, thereby avoiding trial and error and enabling advanced resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a UE uses trial and error to determine the time horizon for CSI prediction, then the UE can achieve acceptable prediction accuracy, but the ML model requires repeated re-training and re-configuration, increasing time and computational overhead
Solution Approach 1:
The network node provides feedback to the UE about the predicted time horizon, enabling the UE to adjust its prediction horizon without trial-and-error re-training. The feedback mechanism allows the system to learn from past predictions and optimize the time horizon parameter efficiently, resolving the contradiction between achieving accurate predictions and avoiding repeated re-training overhead.
Solution Approach 2:
The UE autonomously determines its own predicted time horizon based on channel conditions and prediction accuracy requirements, without requiring network-controlled trial and error. This self-service approach allows the UE to independently optimize its prediction horizon, eliminating the need for repeated re-training while maintaining prediction accuracy.
2Adaptability or versatility
If a UE uses variable time horizons for CSI prediction, then the prediction can adapt to different channel conditions, but the radio base station scheduler cannot allocate resources in advance due to inconsistent prediction validity guarantees
Solution Approach 1:
The system changes the time horizon parameter dynamically based on channel conditions and prediction accuracy, allowing the UE to adapt its predictions to varying wireless environments. By providing feedback about the predicted time horizon to the network, the system maintains consistency in resource allocation while preserving adaptability in prediction.
3Productivity
If the network controls the time horizon for UE predictions, then resource allocation can be optimized, but the UE cannot determine the optimal time horizon based on its own ML model performance and channel conditions
Solution Approach 1:
The feedback message acts as an intermediary between the UE's autonomous prediction capability and the network's resource allocation needs. The UE determines the optimal time horizon based on its ML model performance, and the feedback mechanism translates this into network-compatible information, enabling both autonomous determination and optimized resource allocation.
Data Source
AI summary
A method performed by a network node for an adaptive prediction of a time horizon for at least one key performance indicator, KPI, in a wireless network is provided. The method includes transmitting towards a user equipment, UE, a machine learning, ML, model to predict the at least one KPI; and receiving from the UE a first message including the predicted at least one KPI and a predicted time horizon for the predicted at least one KPI. A method performed by a UE is also provided.