The generalized variable step-size diffusion continuous mixed p-norm (GVSS-DCMPN) algorithm is proposed in this paper, which is derived based on the improved continuous mixed p-norm (CMPN) strategy. In detail, a linear function is designed for the CMPN strategy. The proposed GVSS-DCMPN algorithm capable of exploiting various error norms to obtain performance improvement in non-Gaussian noise environment can be viewed as the generalization of the traditional p-norm-based algorithms in the sense of continuous errors. In particular, the GVSS-DCMPN algorithm transforms into the VSS-DCMPN algorithm when the slope of the linear function is set to 0. The computational complexity as well as the mean convergence is analyzed in the paper. Simulation results over the diffusion network show that the proposed algorithm achieves performance gain over some existing diffusion algorithms.

提出了基于改进的连续混合p-范数（CMPN）策略的广义变步长扩散连续p-范数（GVSS-DCMPN）算法。详细地，线性函数被设计用于CMPN策略。提出的GVSS-DCMPN算法能够利用各种误差准则来提高非高斯噪声环境下的性能，可以看作是传统p的推广。连续错误意义上基于-norm的算法。特别是，当线性函数的斜率设置为0时，GVSS-DCMPN算法将转换为VSS-DCMPN算法。本文分析了计算复杂度和均值收敛性。在扩散网络上的仿真结果表明，与现有的一些扩散算法相比，该算法具有更高的性能。