Coded Digital Predistortion for Phased Arrays With PA Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital predistortion (DPD) techniques for radio frequency (RF) systems with phased antenna arrays face challenges in differentiating between contributions from individual power amplifiers (PAs), leading to inefficiencies and increased complexity, particularly in millimeter-wave/5G technologies.
Innovation Solution
A coded DPD arrangement that applies orthogonal or non-orthogonal codes to input signals for each PA in a phased antenna array, using a coded beamformer circuit and DPD adaptation circuit to generate and update predistortion coefficients based on feedback signals, allowing for improved differentiation between PA contributions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional DPD techniques are used for phased antenna arrays, then the implementation is simpler, but the ability to differentiate between individual PA contributions is poor and complexity increases
Solution Approach 1:
The patent introduces coded sequences as an intermediary tool to differentiate PA contributions. Each PA is assigned a unique coded sequence that modulates its output signal. These codes act as identifiers that allow the feedback signal to be decomposed and attributed to specific PAs, enabling precise measurement of individual PA contributions without requiring direct physical separation or complex individual monitoring circuits for each PA.
2Manufacturing precision
If DPD is applied to enhance PA linearity, then modulation accuracy improves, but the system complexity and computational requirements increase
Solution Approach 1:
The patent segments the DPD system by assigning unique coded sequences to each PA group or individual PA. This segmentation allows the overall DPD problem to be broken down into smaller, independent sub-problems that can be processed separately. The feedback signal can be decomposed using correlation with the known codes, enabling independent characterization and compensation of each PA's nonlinearities without requiring a monolithic complex model of the entire array.
Solution Approach 2:
The patent changes the parameter space by introducing coded sequences as an additional dimension for signal differentiation. Instead of relying solely on spatial or temporal separation methods, the system modifies the signal parameters by embedding unique codes in each PA's output. This parameter change enables more efficient identification and compensation of PA nonlinearities through correlation-based processing, reducing the computational complexity compared to traditional methods.
3Reliability
If feedback from PA output is used to update DPD model, then predistortion effectiveness improves, but the time required for DPD adaptation increases
Solution Approach 1:
The patent employs periodic coded sequences that are transmitted in a time-division manner. Each code is activated for a specific time interval, allowing the system to collect feedback signals corresponding to each PA's contribution during its designated time slot. This periodic activation pattern enables the DPD model to be updated using feedback from all PAs in a systematic and efficient manner, reducing the total adaptation time compared to sequential testing methods while maintaining comprehensive coverage of all PA contributions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Digital predistortion (DPD) arrangements that use coding to differentiate between contributions from individual power amplifiers (PAs) is disclosed. When used with a phased antenna array having N antenna elements and N corresponding PAs, the arrangements may sequentially apply N codes to input signals provided to N PAs. Each code is a vector having N elements, where each element corresponds to a different PA in that the element is a complex gain applied to an input signal for the PA. Together, the N codes may be arranged as N rows or N columns of a matrix P. The disclosed arrangements may generate N feedback signals by applying the matrix P to a signal generated by a probe antenna element sensing wireless RF signals transmitted by the N antenna elements, and to update DPD coefficients based on the N feedback signals. The N codes may be either orthogonal or non-orthogonal.