RNN Self-Interference Cancellation for Multi-Band Full Duplex Radios

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Wireless communication systems face challenges in managing self-interference noise, particularly in full duplex communication systems using multiple frequency bands, which affects the accuracy and efficiency of 5G wireless networks, especially in remote settings with environmental degradations and interference from co-located antennas and nonlinear power amplifiers.

Innovation Solution

Utilizing recurrent neural networks to generate adjusted signals that compensate for self-interference noise by modeling higher-order memory effects of interference signals, including harmonics and intermodulation components, thereby enabling full duplex transmission across multiple frequency bands.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If full duplex communication is implemented using multiple frequency bands, then spectrum utilization and network capacity are improved, but self-interference noise from co-located antennas and nonlinear power amplifiers increases

Engineering Contradiction:

Improvenetwork capacityVSAvoidself-interference noise

Core Design Contradiction:

ProductivityVSObject-generated harmful factors

Solution Approach 1:

The patent models the self-interference noise characteristics (harmonics and intermodulation components) using recurrent neural networks to convert the harmful interference into predictable patterns that can be compensated for and removed, thereby enabling full duplex communication on multiple frequency bands

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

2Measurement precision

If recurrent neural networks are used to model higher-order memory effects of interference signals, then self-interference noise cancellation accuracy is improved, but device complexity increases

Engineering Contradiction:

Improvenoise cancellation accuracyVSAvoidcomputational complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent replaces traditional mechanical or circuit-based filtering methods with recurrent neural network-based digital signal processing to model and cancel self-interference noise, achieving higher accuracy through software-based adaptive filtering

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Adaptability or versatility

If multiple frequency bands are used for simultaneous transmission and reception, then spectrum utilization is improved, but interference management becomes more difficult

Engineering Contradiction:

Improvespectrum utilizationVSAvoidinterference management complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent employs recurrent neural networks that process historical interference data and provide feedback to continuously refine the self-interference model, enabling adaptive cancellation across multiple frequency bands and simplifying interference management through learned patterns

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12489478B2Self interference noise cancellation to support multiple frequency bands with neural networks or recurrent neural networks

Publication Date: 2025.12.02 MICRON TECHNOLOGY INC

AI summary

Examples described herein include systems and methods which include wireless devices and systems with examples of multiple frequency bands transmission with a recurrent neural network that compensates for the self-interference noise generated by power amplifiers at harmonic frequencies of a respective wireless receiver. The recurrent neural network may be coupled to antennas of a wireless device and configured to generate the adjusted signals that compensate self-interference. The recurrent neural network nay include a network of processing elements configured to combine transmission signals into sets of intermediate results. Each set of intermediate results may be summed in the recurrent neural network to generate a corresponding adjusted signal. The adjusted signal is receivable by a corresponding wireless receiver to compensate for the self-interference noise generated by a wireless transmitter transmitting on the same or different frequency band as the wireless receiver is receiving.