Search

Because of the inherent nonlinearity of high power amplifier, there are in-band distortion and adjacent-channel interference, which may have a negative influence on communication systems. We have to make a linearization processing to overcome them. First, this paper attempts to analyze nonlinearity distortion of HPA in a mathematic approach, and then briefly introduces the basic principle about digital predistortion. Furthermore, after depicting simple neural network, the paper proposed an adaptive digital predistortion technology based on Single Input and Single Onput Neural Network(SISO-Neural Network), which can improve third-order and fifth-order intermodulation effectively. The proposed technique is more advantageous on convergence speed and complicacy of hardware realization compared with polynomial-fitting technique. Last, a matlab simulation of double-tone signal is done and the results also verify that the proposed technique is superior.

A new recognition algorithm to the symbol shaped signals is presented, in which eight digital modulation types are classified. The infection to signals and parameterspsila extraction aroused by symbol shape is introduced in detailed, and the modulation types are classified from the decision-tree classifier. The computer simulation results show that, the algorithm has a good recognition performance for the shaped signals, further more it has a better recognition performance than the methods without consideration of the symbol shape.