Shared TDD Receiver Path for RNN Power Amplifier Linearization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in efficiently compensating for nonlinear power amplifier noise, which affects signal quality and requires specialized hardware for digital pre-distortion filters, leading to complexity and resource inefficiency.
Innovation Solution
A recurrent neural network is used in conjunction with a digital pre-distortion filter to compensate for nonlinear power amplifier noise, utilizing a time division duplexing configuration to share a single receiver path for both feedback and signal reception, optimizing computational resources and board space by activating a switch to provide feedback during uplink time periods and deactivating it during downlink periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a digital pre-distortion filter is used to compensate for nonlinear power amplifier noise, then signal quality is improved, but device complexity increases due to requiring specialized hardware
Solution Approach 1:
The patent replaces traditional hardware-based digital pre-distortion filters with a software-based recurrent neural network implementation. The RNN is executed on general-purpose processors or FPGAs, eliminating the need for specialized analog or digital circuitry. This substitution maintains signal quality improvement while reducing hardware complexity and resource requirements.
Solution Approach 2:
The recurrent neural network is designed to perform multiple functions: it compensates for power amplifier nonlinearities, adapts to changing channel conditions, and can be reconfigured for different operating scenarios. This multi-functionality replaces what would traditionally require multiple specialized hardware components, thereby reducing overall device complexity while maintaining or improving signal quality.
2Productivity
If separate hardware paths are used for feedback and signal reception, then processing capability is improved, but board space and resource efficiency deteriorate
Solution Approach 1:
The patent merges the feedback path and signal reception path into a single shared hardware path. A switch dynamically routes signals through this shared path: during uplink periods, the path carries feedback signals for RNN training; during downlink periods, it carries received signals. This time-division multiplexing approach maintains full processing capability while halving the board space required for dedicated paths.
Solution Approach 2:
The system employs periodic time-division duplexing where the shared receiver path alternates between handling feedback signals and handling received signals. During uplink time periods, the path processes feedback for RNN adaptation; during downlink time periods, it processes incoming signals. This periodic switching maintains processing capability for both functions while using a single physical path, thereby reducing board space.
3Measurement precision
If specialized hardware is implemented for each signal processing function, then processing precision is improved, but ease of manufacture and adaptability deteriorate
Solution Approach 1:
The patent replaces specialized hardware implementations with software-based processing using recurrent neural networks. The RNN can be deployed on standard processors, DSPs, or FPGAs, eliminating the need for custom-designed hardware circuits. This substitution maintains processing precision through software algorithms while dramatically simplifying manufacturing, as no specialized hardware design or fabrication is required.
Solution Approach 2:
The system achieves different processing functions by changing software parameters and model configurations rather than hardware design. The recurrent neural network can be reconfigured through parameter adjustments to adapt to different power amplifier characteristics, frequency bands, and operating conditions. This parameter-based adaptability simplifies manufacturing while maintaining processing precision across various scenarios.
Data Source
AI summary
Examples described herein include methods, devices, and systems which may compensate input data for nonlinear power amplifier noise to generate compensated input data. In compensating the noise, during an uplink transmission time interval (TTI), a switch path is activated to provide amplified input data to a receiver stage including a recurrent neural network (RNN). The RNN may calculate an error representative of the noise based partly on the input signal to be transmitted and a feedback signal to generate filter coefficient data associated with the power amplifier noise. The feedback signal is provided, after processing through the receiver, to the RNN. During an uplink TTI, the amplified input data may also be transmitted as the RF wireless transmission via an RF antenna. During a downlink TTI, the switch path may be deactivated and the receiver stage may receive an additional RF wireless transmission to be processed in the receiver stage.