Cross-Band Spatial Filter Prediction for Faster Beam Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In communications systems, measuring spatial filters for multiple frequency bands increases overheads and measurement delays for terminal devices, particularly in systems like 5G and NR, due to the need to traverse all combinations of transmit and receive beams.
Innovation Solution
A spatial filter prediction method that utilizes neural networks, such as CNNs and LSTMs, to predict spatial filters for unmeasured frequency bands based on measurements from other bands, reducing the need for comprehensive beam sweeping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the terminal device measures all spatial filters corresponding to multiple frequency bands, then the spatial filter quality is improved, but the overheads and measurement delay increase
Solution Approach 1:
The terminal device performs measurement on spatial filters in a first frequency band in advance, and uses these preliminary measurement results to predict spatial filter qualities in a second frequency band, avoiding the need to measure all spatial filters in all frequency bands before selection
Solution Approach 2:
A prediction model acts as an intermediary between the measurement results in the first frequency band and the spatial filter selection in the second frequency band, enabling the system to infer unmeasured spatial filter qualities without direct measurement
2Measurement precision
If the terminal device measures all spatial filters corresponding to multiple frequency bands, then the spatial filter quality is improved, but the overheads and measurement delay increase
Solution Approach 1:
The patent extracts only the essential measurement data from the first frequency band that is sufficient for prediction, removing the need to transmit and process all measurement results from all frequency bands, thereby reducing overheads while maintaining prediction accuracy
3Measurement precision
If the terminal device performs comprehensive beam sweeping for spatial filter measurement, then the measurement precision is improved, but the complexity of the device increases
Solution Approach 1:
Comprehensive beam sweeping and measurement are performed in advance on a subset of frequency bands, and the results are stored for later prediction, eliminating the need to repeat the complex beam sweeping process for all frequency bands
Solution Approach 2:
The patent creates a predictive copy of the spatial filter quality information from measured frequency bands to unmeasured frequency bands using a prediction model, avoiding the need to perform actual measurements and complex beam sweeping in all bands
Data Source
Figure 1~2
Figure 3
Figure 4(a)~4(c)
AI summary
The present application provides spatial filter prediction methods, a terminal device, and a network device. A method comprises: a terminal device measures a first spatial filter set; and, according to a measurement result of the first spatial filter set, the terminal device performs spatial filter prediction on a second spatial filter set, wherein the first spatial filter set corresponds to a first frequency band, and the second spatial filter set corresponds to a second frequency band.