ADAPTIVE LINEARIZATION OF EFFICIENT HIGH POWER AMPLIFIERS USING POLYNOMIAL PREDISTORTION WITH GLOBAL OPTIMIZATION
| dc.contributor.advisor | Kumar, S. | |
| dc.creator | Ghaderi, Mohammad | |
| dc.date.accessioned | 2018-12-07T21:59:54Z | |
| dc.date.available | 2018-12-07T21:59:54Z | |
| dc.date.issued | 1994-11 | |
| dc.date.submitted | November 1994 | en_US |
| dc.description.abstract | Analog frequency modulation (FM) or nearly constant envelope digital modulation schemes are used for most of the present generation mobile radio systems. Because of the almost constant envelope property, a high power-efficient nonlinear amplifier can be used for such systems. However, since high demand for cellular systems will lead to spectrum congestion, spectrally efficient linear modulation methods such as quadrature amplitude modulation (QAM) have been proposed for the future mobile radio systems. As these modulation methods are linear in nature, amplification must also be linear. Linear power amplifiers have poor power efficiency and are unsuitable for mobile transceivers where battery power must be conserved. To achieve both linearity and power efficiency, an efficient high power nonlinear amplifier must be linearized; predistortion is an effective technique for linearizing transmitter power amplifiers. Two types of predistortion linearizers, namely, the lookup table predistorter, and the polynomial predistorter, have been reported in the literature. Polynomial predistorters have an advantage over lookup table predistorters, however, those reported in the past are limited to a polynomial of the fifth order, because of optimization problems, and only power amplifiers with nonlinearities of fifth order or less can be linearized using such linearizers. Different types of polynomial predistortion linearizers are proposed in this thesis. Important features of these linearizers are: objective functions of quadratic shape are used for optimization, circuit imperfections can be compensated adaptively, the order of the predistorter is not limited to the fifth, and they are not limited to power amplifiers with fifth order nonlinearities. First, an adaptive linearizer with amplitude and phase control circuits is introduced. The amplitude and phase control signals are generated using gain and phase polynomial functions. The power amplifier input and output signals are demodulated and used for optimizing the predistorter polynomial coefficients. The recursive least squares algorithm is employed for the optimization process. This linearizer, however, does not compensate for the predistorter and demodulator imperfections. A modified version of this linearizer is proposed which can compensate for imperfections adaptively. In this linearizer, the power amplifier output signal is demodulated using the predistorter input signal and the predistorter and demodulator imperfections are estimated by sampling the demodulated signals after updating the predistorter coefficients. The results of computer simulation show that the proposed linearizers are effective in compensating the amplifier nonlinearities, even in the presence of circuit imperfections. Spectrum spreading improvement of 52 dB is achieved at the edge of the filtered band after 60 iterations. Next, an adaptive predistortion linearizer with a complex quadrature modulator is proposed. This linearizer has less circuit complexity than the one with amplitude and phase control circuits. A polynomial postdistorter is employed in the adaptation algorithm for optimizing the predistorter coefficients. Using this linearizer , spectrum spreading is improved by 50 dB at the filtered band edge after 600 iterations, in the presence of circuit imperfections. A modified version of this linearizer with convergence faster by an order of magnitude, is proposed. It is shown that the spectral performance of the polynomial linearizer degrades rapidly near saturation. A new nonlinear function for the predistorter gain functions is proposed. Using this function, the power amplifier is linearized almost up to its saturated output power, and out-of-band power spectral density of 88 dB below that at the bandpass center frequency is achieved. | en_US |
| dc.identifier.uri | http://hdl.handle.net/10388/11615 | |
| dc.title | ADAPTIVE LINEARIZATION OF EFFICIENT HIGH POWER AMPLIFIERS USING POLYNOMIAL PREDISTORTION WITH GLOBAL OPTIMIZATION | en_US |
| dc.type.genre | Thesis | en_US |
| thesis.degree.department | Electrical and Computer Engineering | en_US |
| thesis.degree.discipline | Electrical Engineering | en_US |
| thesis.degree.grantor | University of Saskatchewan | en_US |
| thesis.degree.level | Doctoral | en_US |
| thesis.degree.name | Doctor of Philosophy (Ph.D.) | en_US |