DPT for Linear and Nonlinear Distortion - Part I — Lesson 5

This lesson covers the digital predistortion techniques for linear and non-linear distortions. The lesson explains the selection of topology and algorithms, the application of filtering algorithms, and the use of linear filter theory. It also discusses the Wiener Hammerstein model, the use of FIR filters, and the mapping of non-linear to linear terms. The lesson further explores different types of processing, such as batch mode and sample to sample processing, and various methods like least square, gradient descent, and recursively least square methods. The lesson concludes with a practical example of a Doherty power amplifier and the effect of IQ imbalances.

Video Highlights

00:18 - Linear filter theory and the use of FIR filter
13:58 - Least square methods for batch mode processing
16:30 - Gradient descent method and mean square algorithms
18:52 - Least mean square method
26:25 - Inverse modeling performance
28:52 - Effect if IQ imbalance on the transmitter output

Key Takeaways

- Digital predistortion techniques are used in software-defined radios to correct distortions.
- The selection of topology and algorithms is crucial in the modeling process.
- Linear and non-linear techniques, along with various algorithms, can be applied to achieve desired results.
- Different types of processing, such as batch mode and sample to sample processing, have their own advantages and complexities.
- Practical examples, like the Doherty power amplifier, help in understanding the application of these techniques.
- IQ imbalances can cause additional distortions, which can be corrected using digital predistortion techniques.