A momentum fractional order multimodulus blind equalization algorithm

Abstract

The multimodulus algorithm (MMA) is widely used to suppress inter-symbols interference (ISI) in communication systems because of its advantages that it can complete blind equalization and carrier phase recovery simultaneously. However, the slow convergence speed of MMA causes the equalizer to be unable to eliminate ISI effectively, which deteriorate the communication quality of the system. For that, we first present a momentum fractional order MMA (MFOMMA) to improve the performance of MMA for blind equalization in this paper. By optimizing MMA with the momentum term and the fractional order gradient information of the cost function, the proposed algorithm can get better ISI suppression performance. Then, the convergence performance of MFOMMA is analyzed to reveal the relationship between algorithm parameters and equalization ability. Moreover, in order to further enhance the ISI elimination capability of MFOMMA, we develop a variable fractional order scheme, which guarantee that the algorithm not only obtains a faster convergence speed but also a lower residual ISI. Finally, the effectiveness and superiorities of the proposed algorithm are demonstrated by some simulation experiments.