RNN Power Amplifier Compensation Using a Shared TDD Receiver Path

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional wireless devices face complexity and inefficiency in implementing digital pre-distortion (DPD) filters to compensate for nonlinear power amplifier noise, often requiring specialized hardware and separate paths for feedback processing, which consumes computational resources and board space.

Innovation Solution

A recurrent neural network (RNN) is used in conjunction with a time division duplexing (TDD) configured radio frame to provide both feedback and wireless transmission signals through a single receiver path, activating a switch to route feedback during uplink periods and deactivating it during downlink periods, allowing efficient use of resources and optimizing board space.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If separate feedback processing paths are used for DPD filter compensation, then noise compensation accuracy is improved, but device complexity and board space increase

Engineering Contradiction:
Improvenoise compensation accuracyVSAvoiddevice complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent combines the feedback processing path with the existing receiver path into a single integrated path. The switch dynamically routes signals to share hardware resources between DPD feedback processing and regular signal reception, eliminating the need for separate dedicated feedback processing hardware while maintaining compensation accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The receiver path is designed to serve dual purposes: processing DPD feedback signals during uplink periods and receiving wireless transmission signals during downlink periods. This multi-functional design allows the same hardware components to be reused for different functions at different times, reducing overall device complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Productivity

If dedicated hardware is implemented for DPD filter processing, then processing capability is improved, but board space and manufacturing complexity increase

Engineering Contradiction:
Improveprocessing capabilityVSAvoidmanufacturing complexity
Core Design Contradiction:
ProductivityVSEase of manufacture

Solution Approach 1:

The system employs a dynamic switch that can reconfigure the receiver path between different functions based on operational mode (uplink/downlink). This dynamic reconfiguration allows the same hardware to adapt to different processing requirements without needing separate dedicated hardware for each function, simplifying manufacturing.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The existing receiver path infrastructure is utilized to process DPD feedback signals, allowing the system to serve itself rather than requiring additional dedicated hardware. The receiver path components (amplifiers, filters, ADCs) are already present and can be repurposed for feedback processing during uplink periods.

Inventive Principle:
Principle #25Self-service

3Reliability

If computational resources are allocated for separate feedback processing, then noise compensation performance is improved, but computational resource usage increases

Engineering Contradiction:
Improvenoise compensation performanceVSAvoidcomputational resource usage
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system uses periodic time-division multiplexing where the receiver path alternates between processing DPD feedback during uplink periods and receiving wireless signals during downlink periods. This periodic sharing of computational resources maintains noise compensation performance during uplink while reducing overall resource usage compared to continuous dedicated processing.

Inventive Principle:
Principle #19Periodic action

Data Source

PatentUS11601146B2Wireless devices and systems including examples of compensating power amplifier noise with neural networks or recurrent neural networks
Publication Date: 2023.03.07 LODESTAR LICENSING GROUP LLC
  • US11601146B2 patent drawing
  • US11601146B2 patent drawing
  • US11601146B2 patent drawing

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.