Spectrum Occupancy Prediction Using Multi-Dimensional Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing spectrum occupancy prediction methods in wireless communication systems, particularly in 5G and beyond, fail to effectively address multi-dimensional correlations in time, frequency, and space, leading to high computational complexity and the need for continuous data availability, which is not always feasible due to security threats or natural disasters.
Innovation Solution
A trainable model architecture comprising first and second types of trainable modules for time/frequency and spatial correlations, respectively, allowing for re-trainability and reduced complexity by omitting inactive network elements, with an end-classifier for final decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multi-dimensional correlation (time, frequency, space) is exploited for spectrum occupancy prediction, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments the spectrum occupancy prediction task into multiple independent trainable modules, where each module handles prediction for a specific communication device based on its past occupancy data. This segmentation allows parallel processing of multiple devices' predictions, reducing overall computational complexity while maintaining the ability to exploit multi-dimensional correlations within each module's localized data scope.
2Reliability
If continuous data availability is required for accurate prediction, then prediction reliability is improved, but system vulnerability increases due to security threats or natural disasters
Solution Approach 1:
The patent implements preliminary action by training individual device-specific models offline using historical occupancy data before actual prediction is needed. These pre-trained models can then operate independently with minimal real-time data requirements, allowing the system to maintain prediction functionality even when continuous data availability is compromised by security threats or natural disasters.
3Adaptability or versatility
If individual models are trained for each communication device, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent applies local quality by training each device's prediction module using only that device's own past occupancy data, creating localized models adapted to each device's specific characteristics and environment. This approach improves adaptability to individual device patterns while reducing complexity compared to training a single global model on all devices' data, as each module processes only relevant local information.
Data Source
AI summary
The present disclosure relates to a spectrum occupancy prediction, which employs a first trainable module and a second trainable module. The first and the second trainable modules predict spectrum occupancy at a respective first and second communication devices based on past occupancies and/or occupancies in adjacent subband(s). The prediction from the first trainable module and the second trainable module is input to a trainable output (third) module, which then provides the spectrum occupancy prediction.