Grouped MIMO DPD for PA Linearization With Lower Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Advanced Antenna Systems (AAS) with multiple input multiple output (MIMO) transmitters face challenges in reducing hardware complexity and addressing non-linear distortions caused by Power Amplifiers (PAs), especially in isolator-free configurations, where existing digital predistortion (DPD) techniques suffer from high computational complexity and are not universally applicable for correlated and uncorrelated signals.
Innovation Solution
The implementation of an Iterative Learning Control (ILC) combined MIMO DPD scheme with kernel regression, which groups antenna branches to reduce computational and implementation complexity, and is independent of signal correlation, effectively predistorting input signals to linearize PA outputs across multiple branches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional DPD techniques are used in isolator-free AAS, then PA non-linear distortion is compensated, but computational complexity increases significantly
Solution Approach 1:
The patent segments the MIMO system into multiple independent SISO DPD chains, where each transmit branch has its own DPD unit that processes only its own input signal. This segmentation eliminates the need for complex MIMO-specific DPD algorithms while maintaining linearization performance, as each SISO DPD independently compensates for PA non-linearities without requiring computation of inter-branch interference terms.
Solution Approach 2:
The patent employs a universal SISO DPD architecture that can be applied to each transmit branch independently, making the solution applicable to both correlated and uncorrelated signal scenarios. This universal approach avoids the need for separate MIMO DPD algorithms and works effectively across different signal correlation conditions, reducing overall system complexity.
2Device complexity
If hardware components like RF isolators are removed to reduce complexity, then hardware structure is simplified, but PA protection from reflected signals is lost and non-linear distortion increases
Solution Approach 1:
The patent replaces the mechanical RF isolator with a digital signal processing solution. Instead of using physical isolation components to protect PAs from reflected signals, the system uses SISO DPD algorithms that digitally compensate for the effects of antenna mismatch and mutual coupling, thereby maintaining PA protection and linearization performance without additional hardware complexity.
3Reliability
If per-branch DPD is implemented in MIMO systems, then each branch is linearized independently, but computational resources and power consumption increase
Solution Approach 1:
The patent segments the DPD function into independent per-branch SISO processors, where each branch's DPD unit only needs to process its own input signal and estimate its own PA characteristics. This segmentation reduces the overall computational burden compared to joint MIMO DPD, as each SISO DPD operates independently without requiring computation of inter-branch interference terms, thereby reducing power consumption while maintaining linearization accuracy.
Data Source
AI summary
Systems and methods are disclosed herein for Digital Predistortion (DPD) in a Multiple Input Multiple Output (MIMO) transmitter. In some embodiments, a MIMO transmitter comprises a plurality of antenna branches comprising a respective plurality of power amplifiers coupled to a respective plurality of antenna elements. The MIMO transmitter also includes one or more DPD systems operable to predistort one or more respective groups of input signals to provide one or more respective groups of predistorted input signals for one or more respective groups of antenna branches. Each group of antenna branches comprises at least two of the plurality of antenna branches. In some embodiments, the MIMO transmitter is a massive MIMO transmitter. Embodiments of a per-branch DPD scheme for a MIMO transmitter that uses Iterative Learning Control (ILC) and kernel regression are also disclosed.