The Research Of Distributed Adaptive Filtering Algorithm Based On Generalized Maximum Correntropy Criterion

At present,the distributed adaptive algorithm in most cases is based on the minimum mean square error criterion,it can show good performance in dealing with Gaussian noise,but in the impulsive noise environment,the performance will deteriorate.In this paper,we propose a distributed maximum correntropy criterion algorithm for impulsive noise,DMCC can not achieve faster convergence speed and lower steady-state error at the same time when using fixed step size.Based on the variable step size strategy of maximum correlation entropy criterion,a distributed maximum correntropy criterion variable step size algorithm(DMCC-VSS)is proposed.Through the analysis of the computational complexity of the algorithm,we can see that the computational complexity of the algorithm is equivalent to that of the algorithm based on the mean square criterion,and is lower than the algorithm based on the minimum residual entropy criterion.By analyzing the mean and mean square performance of the algorithm,the stability condition of the algorithm is also given.In this paper,two groups of simulation experiments are set up for DMCC-VSS algorithm.The simulation results show that the algorithm has smaller steady-state error than the variable step size algorithm of mean square criterion in the same impulsive environment,faster convergence speed than DMCC algorithm,and better convergence in different probability impulsive noise environment.In addition,in many important applications,such as regression and pattern classification,there may be some differences between tasks.At this time,single task network is no longer applicable,and multi task network needs to be established to complete such tasks.In the algorithm of multi task network,there is still a lack of adaptive filter with better anti impulse noise performance.In this paper,the generalized maximum correlation entropy(GMCC)criterion is applied to the multi task distributed network,and a multi task distributed adaptive filtering algorithm based on GMCC is proposed.Similarly,the mean and mean square performance of the algorithm are analyzed,the boundary range of step convergence is given,and its stability is analyzed.In this paper,we set up four groups of simulation experiments on the algorithm.The simulation results show that the algorithm has better convergence performance than other algorithms in the case of Gaussian and non Gaussian,impulse noise environment with different probabilities,estimation task correlation and task cooperation and non cooperation strategy scenarios.