Modeling based on Elman wavelet neural network for class-D power amplifiers

摘要

In Class-D Power Amplifiers (CDPAs), the powersupply noise can intermodulate with the input signal, manifestinginto power-supply induced intermodulation distortion (PS-IMD)and due to the memory effects of the system, there existasymmetries in the PS-IMDs. In this paper, a new behavioralmodeling based on the Elman Wavelet Neural Network (EWNN)is proposed to study the nonlinear distortion of the CDPAs.In EWNN model, the Morlet wavelet functions are employedas the activation function and there is a normalized operationin the hidden layer, the modification of the scale factor andtranslation factor in the wavelet functions are ignored to avoidthe fluctuations of the error curves. When there are 30 neuronsin the hidden layer, to achieve the same square sum error(SSE) emin = 10-3, EWNN needs 31 iteration steps, while thebasic Elman neural network (BENN) model needs 86 steps. TheVolterra-Laguerre model has 605 parameters to be estimatedbut still can’t achieve the same magnitude accuracy of EWNN.Simulation results show that the proposed approach of EWNNmodel has fewer parameters and higher accuracy than theVolterra-Laguerre model and its convergence rate is much fasterthan the BENN model.