Wideband Power Amplifier Predistortion With Stable Direct Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In wireless mobile communication systems, direct learning algorithms for power amplifiers face instability and inefficiency, particularly in wideband signals, due to numerical error accumulation and the need for frequent restarts, which violates 3GPP specifications and leads to increased inter-modulation products.
Innovation Solution
An adaptive direct learning algorithm using a conjugate gradient method is implemented, where the initial residual and direction are set based on previous adaptations, and an error scaling factor is introduced to reduce numerical errors, stabilizing the algorithm and achieving convergence with fewer iterative steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If a direct learning algorithm is used for power amplifier linearization, then the algorithm can operate in compressed regions to improve power efficiency, but numerical error accumulation causes instability and requires frequent restarts
Solution Approach 1:
The patent implements a feedback mechanism where the conjugate gradient algorithm continuously adapts the pre-distorter parameters based on the error between the actual and desired output signals. This feedback loop allows the system to correct numerical errors progressively while maintaining operation in compressed regions, thus improving power efficiency without sacrificing stability.
Solution Approach 2:
The patent applies preliminary action by pre-calculating and storing the conjugate gradient search directions and step sizes before the actual adaptation process. This preparation allows the algorithm to converge faster and reduces the accumulation of numerical errors during runtime, maintaining stability while operating in efficient compressed regions.
2Device complexity
If the conjugate gradient algorithm uses fixed initial residual and direction, then the implementation is simpler, but convergence requires more iterative steps
Solution Approach 1:
The patent makes the initial residual and search direction dynamic by updating them based on previous iteration results. Specifically, the initial residual is set to the current error signal, and the search direction is updated using the conjugate gradient formula incorporating previous direction and current gradient information. This dynamic approach accelerates convergence while maintaining manageable implementation complexity.
Solution Approach 2:
The patent extends the algorithm by incorporating memory of previous iterations into the current adaptation process. By utilizing historical information from previous conjugate gradient steps, the algorithm operates in an extended dimensional space that includes temporal dependencies, enabling faster convergence without significantly increasing implementation complexity.
3Reliability
If the algorithm is restarted frequently to maintain stability, then numerical errors are reset, but this increases inter-modulation products and violates 3GPP specifications
Solution Approach 1:
The patent ensures continuous adaptation of the pre-distorter parameters through the conjugate gradient algorithm without interrupting the signal processing chain. By maintaining continuous operation and using proper numerical techniques to prevent error accumulation, the system avoids restarts that would cause inter-modulation products, thus complying with 3GPP specifications while maintaining numerical stability.
Solution Approach 2:
The patent converts the potentially harmful effect of numerical error accumulation into a beneficial adaptive process. By using the conjugate gradient algorithm's iterative nature to systematically reduce errors rather than resetting through restarts, the system transforms what would be a stability problem into a controlled adaptation mechanism that maintains both stability and spectral compliance.
Data Source
AI summary
The present invention addresses method, apparatus and computer program product for stabilization of the direct learning algorithm for wideband signals. Thereby, a signal to be amplified is input to a pre-distorter provided for compensating for non-linearity of the power amplifier, and the pre-distorted output signal from the pre-distorter is forwarded to the power amplifier. Parameters of the pre-distorter are adapted based on an error between the linearized signal output from the power amplifier and the signal to be amplified using an adaptive direct learning algorithm, and the linear system of equations formed by the direct learning algorithm are solved using a conjugate gradient algorithm, wherein, once per direct learning algorithm adaptation, at least one of the initial residual and the initial direction of the conjugate gradient algorithm are set based on the result of the previous adaptation.