Intermodulation Interference Estimation Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face interference from intermodulation products generated by nonlinear transmission characteristics, which affect signal quality and system performance, especially in high-power broadband multi-standard multicarrier FDD systems.
Innovation Solution
A method using machine learned parameters to weight and combine signals, estimating interference signals caused by intermodulation products, and correcting received signals by subtracting these estimates, thereby improving signal accuracy and reducing memory usage in estimation circuitry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to estimate and cancel intermodulation interference, then signal quality may be maintained, but memory usage in estimation circuitry becomes excessive
Solution Approach 1:
The patent segments the estimation process into two distinct sets of machine learned parameters: a first set for weighting individual received signals before combining, and a second set for weighting the composite signal after combination. This segmentation allows the estimation circuitry to process signals in stages, reducing the memory burden while maintaining estimation accuracy for intermodulation interference cancellation.
2Productivity
If system configuration is optimized for high-power broadband multi-standard multicarrier FDD systems, then system performance improves, but interference from intermodulation products increases
Solution Approach 1:
The patent implements a feedback mechanism where the estimation circuitry continuously estimates the intermodulation interference signal based on received signals and machine learned parameters, then feeds this estimation back to cancel the interference. This closed-loop approach allows the system to dynamically compensate for intermodulation products generated by high-power transmission, maintaining system performance while mitigating harmful interference effects.
Solution Approach 2:
The patent converts the harmful intermodulation interference into a beneficial signal by estimating the interference characteristics using machine learned parameters and then using this estimation to cancel the interference. The harmful intermodulation products are transformed into useful information about the interference pattern, which is then used to improve signal quality through cancellation.
3Measurement precision
If machine learned parameters are used to weight and combine signals, then interference estimation accuracy improves, but device complexity increases
Solution Approach 1:
The patent changes the parameters used for signal processing by introducing machine learned parameters that are optimized for interference estimation. Instead of using traditional fixed weighting coefficients, the system employs learnable parameters that adapt to the specific characteristics of the communication environment, improving estimation accuracy while managing complexity through parameter optimization rather than structural complexity.
Data Source
AI summary
The present subject matter relates to a method including receiving a set of signals from transmitters of a communication system, the communication system having a system configuration. A first and second set of machine learned parameters may be provided in accordance with the system configuration. The received signals may be weighted using the first set of machine learned parameters. The weighted signals may be combined to generate a composite signal. The composite signal may be weighted with the second set of machine learned parameters in order to estimate an interference signal that is caused by the set of signals at a receiver of the communication system.