Behavioral modeling of wireless transmitters for distortion mitigation

Doctoral thesis, 2012

As the demand for high speed and reliable wireless communication increases, the importance of having a linear transmitter has enhanced. Distortions created by the transmitter, such as power amplifier nonlinearity and I/Q imbalance, diminish the fidelity and limit the performance of wireless systems, if left undealt with. Among the techniques commonly used to mitigate these distortions, digital predistortion has established itself as a suitable candidate that minimizes the hardware overhead and only requires modest additional power in the transmitter architecture. An important pre-requisite for utilizing digital predistorters is developing accurate and low complex behavioral models, which is the main focus of this thesis. After analyzing the importance of modeling and compensating for the distortion created by modulators and power amplifiers in the transmitter architecture, an overview of some commonly used models in the literature is presented. A novel behavioral modeling approach is proposed which is capable of modeling long term memory effects in power amplifiers, and a new dual–input modeling approach for I/Q imbalance compensation is presented that successfully compensates for distortion created by the modulator. Compared with conventional and recently proposed techniques, the approaches presented in this thesis show promising results in modeling transmitters accurately. The important issue of computational complexity in behavioral models is also discussed, and the accuracy/compexity tradeoff of some common behavioral models is analyzed. Once behavioral modeling techniques are established, they are used for digital predistortion of wireless transmitters. Issues such as identification of digital predistorters and adaptation of parameters due to changes in power amplifier behavior are discussed and a new measurement testbed to evaluate the performance in parameter adaptation algorithms is proposed. The methods and techniques proposed in this work provide ways to both mitigate distortion in and evaluate performance of wireless transmitters in terms of accuracy and complexity, and can help contribute to a better service of quality in wireless communication systems.