Data predistortion with adaptive fuzzy systems

Presents two data predistortion methods for communication transmitters that employ adaptive fuzzy systems. The first is based on indirect learning and does not require a model of the high power amplifier. The second requires an estimated model of the pulse shaping filter and the high power amplifier before training. Simulations with 16 QAM signals show that the performance of fuzzy predistorters are roughly the same as the performance of a third order polynomial predistorter with odd-ordered terms only. Fuzzy and polynomial predistorters in the second method both contain nonlinear parameters, which are sensitive to initial conditions. The fuzzy predistorter has the unique advantage in that the initial values of all adjustable parameters can be set easily based on their linguistic meanings.