Massive Array DPD Training with Iterative Bandwidth Expansion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital predistortion (DPD) techniques for wireless communication systems face challenges in transmitting signals at higher power levels while maintaining compliance with out-of-band (OOB) emissions regulations, often requiring significant power back-off that reduces the overall signal power and efficiency.
Innovation Solution
A DPD training procedure is implemented through an iterative process where the transmitter applies DPD to increasing bandwidth subsets, with feedback from the receiver used to adjust predistortion coefficients, allowing for efficient transmission over maximum allowable bandwidth without increasing power back-off.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If power back-off is applied to decrease OOB emissions below the threshold, then OOB emissions compliance is improved, but the overall power of the transmitted signal is significantly reduced
Solution Approach 1:
The patent applies digital predistortion to the transmitted signal before power amplification to pre-compensate for non-linear distortion. This preliminary action allows the signal to maintain higher power levels while the predistortion coefficients cancel out the OOB emissions that would otherwise be generated, resolving the contradiction between transmitted power and OOB compliance
Solution Approach 2:
The patent dynamically adjusts predistortion coefficients based on feedback from the receiver about actual OOB emissions measurements. By changing these parameters iteratively, the system optimizes the balance between transmitted power and OOB emissions, allowing operation at higher power while maintaining regulatory compliance
2Power
If DPD is applied to the entire maximum allowable bandwidth, then transmission power efficiency is improved, but the complexity of the DPD training process increases
Solution Approach 1:
The patent segments the maximum allowable bandwidth into multiple subsets for iterative DPD training. Instead of training on the full bandwidth at once, the system trains on progressively larger bandwidth portions (e.g., 1/4, 1/2, 3/4, then full bandwidth), which reduces the computational complexity and feedback overhead at each training step while still achieving full-bandwidth DPD performance
Solution Approach 2:
The patent applies partial DPD action by initially training on subsets of the full bandwidth rather than the complete bandwidth. This partial action approach reduces the immediate training complexity and feedback requirements, while the iterative process eventually achieves full-bandwidth predistortion performance
Data Source
Figure 1
Figure 2A~2D
Figure 3
AI summary
This disclosure provides systems, devices, apparatus and methods, including computer programs encoded on storage media, for a DPD training procedure. A base station may transmit, for a plurality of iterations, a signal to at least one UE through a plurality of transmit chains and with application of DPD. The signal transmitted for each iteration may be transmitted with a BW that extends over a plurality of subcarriers and includes pilots extending over a BW subset that increases in subcarrier size for each iteration. The base station may receive, for each iteration, feedback from the at least one UE based on the transmitted signal and apply DPD to each of the plurality of transmit chains based on the feedback. Accordingly, the base station may transmit to one or more UEs through the plurality of transmit chains and with application of the DPD pilot signals that extend over the entire BW.