Power Amplifier Predistortion Update Using Sample Binning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The use of predistortion for power amplifiers increases hardware and computational complexity in transmitters, necessitating an improvement in predistortion techniques to reduce complexity while maintaining performance.
Innovation Solution
An apparatus that allocates samples to bins based on characteristics and determines when to update predistortion, using a threshold-based method to selectively choose samples for updating, thereby reducing the number of samples used for adaptation and decreasing computational complexity, while ensuring accurate predistortion learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If predistortion is applied to power amplifier, then linearization performance is improved, but hardware and computational complexity increases
Solution Approach 1:
The patent segments the sample population by allocating samples to different bins based on their characteristics (e.g., amplitude, phase, or other signal properties). This segmentation allows the predistortion adaptation to be performed selectively on representative samples from each bin rather than all samples, thereby reducing computational complexity while maintaining linearization performance across the entire signal range.
Solution Approach 2:
The patent applies partial action by using only a subset of samples (those allocated to specific bins and meeting threshold criteria) for predistortion adaptation instead of processing all available samples. This selective approach reduces the computational burden while still achieving effective linearization through the representative nature of the selected samples.
2Measurement precision
If all samples are used for predistortion adaptation, then learning accuracy is improved, but computational complexity increases
Solution Approach 1:
Samples are divided into multiple bins based on their characteristics, and only samples from specific bins are selected for predistortion adaptation. This segmentation ensures that the selected samples are representative of different signal conditions while reducing the total number of samples processed, thus maintaining learning accuracy with lower computational complexity.
Solution Approach 2:
The patent changes the parameter of sample selection by introducing bin allocation criteria and threshold-based selection. Instead of using all samples uniformly, the system transforms the sample set based on their characteristic parameters, selecting only those that meet specific bin allocation criteria for adaptation, thereby optimizing the trade-off between learning accuracy and computational complexity.
3Device complexity
If threshold-based sample selection is applied, then computational complexity is reduced, but sample representation quality may deteriorate
Solution Approach 1:
The threshold-based selection is applied within the context of bin segmentation, where samples are first divided into bins based on their characteristics. This ensures that the threshold selection process maintains representativeness across different signal conditions by selecting samples from multiple bins rather than applying a single threshold to all samples, thus preserving sample representation quality while reducing computational complexity.
Solution Approach 2:
Different bins may have different selection criteria or thresholds tailored to their specific characteristic ranges. This local quality approach ensures that each bin's representative samples are selected according to their specific properties, maintaining overall sample representation quality while enabling efficient threshold-based selection across the entire sample set.
Data Source
AI summary
An apparatus is proposed comprising interface circuitry configured to receive a plurality of samples causing an output signal of a power amplifier. The apparatus further comprises processing circuitry configured to allocate at least one sample of the plurality of samples to a bin based on a characteristic of the at least one sample and determine whether a predistortion for samples allocated to the bin is to be updated based on a number of samples allocated to the bin.