ML-Based Positioning Measurement Prediction in 5G Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in accurately predicting future positioning measurements, which affects the efficiency and accuracy of positioning procedures in 5G networks.
Innovation Solution
The implementation of a machine learning model that processes positioning measurements from positioning reference signals (PRS) to predict future positioning measurements, enabling more accurate and efficient positioning procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional positioning measurement methods are used in 5G networks, then positioning procedures can be performed, but signaling overhead increases and power consumption at user equipment increases
Solution Approach 1:
The machine learning model predicts future positioning measurements in advance before they are actually needed, allowing the system to prepare positioning data proactively. This preliminary action reduces the need for frequent real-time measurements and signaling exchanges, thereby lowering power consumption while maintaining positioning accuracy
Solution Approach 2:
Instead of performing actual positioning measurements continuously, the system creates predicted copies of future measurements using the machine learning model. These predicted measurements serve as substitutes for real measurements in many scenarios, reducing the need for actual signal transmissions and receptions at the user equipment
2Measurement precision
If traditional positioning measurement methods are used in 5G networks, then positioning procedures can be performed, but signaling overhead increases
Solution Approach 1:
The machine learning model generates predicted positioning measurements that serve as substitutes for actual measurements. By transmitting these predicted values instead of real-time measurement data, the system significantly reduces signaling overhead while maintaining the necessary positioning information flow
Solution Approach 2:
The invention extracts only the essential positioning information needed for accurate location determination and transmits it, while eliminating redundant signaling. The machine learning model processes raw measurements locally and extracts only the necessary predicted position data for transmission to the network
3Loss of energy
If machine learning model is applied to predict future positioning measurements, then power consumption is reduced and signaling overhead is reduced, but positioning measurement accuracy must be maintained
Solution Approach 1:
The machine learning model is trained using feedback from actual positioning measurements and continuously improves its predictions. The system compares predicted measurements with actual measurements and uses this feedback to refine the model, ensuring that prediction accuracy is maintained while reducing power consumption
Solution Approach 2:
The machine learning model adjusts its internal parameters and prediction algorithms based on changing environmental conditions and movement patterns. By dynamically adapting model parameters, the system maintains high prediction accuracy across different scenarios while minimizing the computational resources required
Data Source
AI summary
Disclosed are techniques for wireless communication. In an aspect, a user equipment (UE) obtains, during a positioning procedure with a location server, one or more positioning measurements of one or more positioning reference signal (PRS) resources transmitted by one or more network nodes, applies a machine learning model to at least the one or more positioning measurements to obtain one or more predicted future positioning measurements associated with the one or more network nodes, and transmits positioning information to the location server, wherein the positioning information comprises the one or more positioning measurements and the one or more predicted future positioning measurements, a current position of the UE determined based, at least in part, on the one or more positioning measurements and a predicted future position of the UE determined based, at least in part, on the one or more predicted future positioning measurements, or both.