Multiband Digital Predistortion for Non-Contiguous Band Linearization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital predistortion (DPD) methods for wideband radio systems face challenges in handling multiple non-contiguous frequency bands, leading to increased complexity and cost, particularly in achieving linear output and meeting 3GPP requirements, as conventional solutions often require higher sampling rates and fail to effectively capture the inverse PA behavior for multiple bands.
Innovation Solution
The method involves performing digital predistortion in two parts: a first part for the full linearization bandwidth using a set of non-linear terms and basis functions, and a second part focusing on the linearization bandwidth excluding the signal's frequency band using another set of non-linear terms and basis functions, allowing for improved error vector magnitude (EVM) and adjacent channel leakage ratio (ACLR) performance, while being computationally efficient and scalable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional DPD methods are used for multiple non-contiguous frequency bands, then linear output and 3GPP requirements can be met, but system complexity and cost increase
Solution Approach 1:
The patent divides the DPD computation into two separate parts: (1) a first DPD part that processes the full linearization bandwidth using a first set of non-linear terms, and (2) a second DPD part that processes only the frequency bands of interest excluding the signal frequency band using a second set of non-linear terms. This segmentation allows the system to meet linear output requirements while reducing the computational complexity compared to conventional single-stage DPD methods that must handle all frequency bands simultaneously.
2Measurement precision
If higher sampling rates are used to handle multiple frequency bands, then signal processing accuracy improves, but computational cost and hardware requirements increase
Solution Approach 1:
The patent extracts and processes only the necessary frequency components in the second DPD part, specifically handling only the frequency bands of interest while excluding the signal frequency band. This selective extraction approach maintains signal processing accuracy for the required bands without the need to process the entire wideband spectrum at high sampling rates, thereby reducing computational cost and energy consumption.
Solution Approach 2:
The first DPD part processes the full linearization bandwidth (excessive action) to ensure all potential intermodulation products are captured, while the second DPD part processes only the specific frequency bands of interest (partial action) to refine the output. This two-stage approach with varying levels of processing detail maintains accuracy where needed while reducing overall computational burden compared to uniform high-rate processing of the entire bandwidth.
Data Source
AI summary
Method and device(s) for performing digital predistortion, “DPD”, on multiple digital input signals to be transmitted in different frequency bands, respectively, of a wireless communication network. The frequency bands being associated with linearization bandwidths, respectively, for application of the DPD. Said DPD is performed in two parts: A first DPD part with a first DPD performed for each of said input signals over the signal's full linearization bandwidth using a first set of non-linear terms and basis functions. A second DPD part where another, second DPD is performed for each of said input signals over the signal's linearization bandwidth except where the linearization bandwidth covers the signal's frequency band using another, second set of non-linear terms and basis functions. Predistorted multiple digital output signals are provided based on both of said performed DPDs.