中文摘要：

二阶Volterra系统二次项将稳定分布的尖峰脉冲特性放大，各阶因素之间耦合强，从而影响了算法的收敛性能。针对传统的Volterra滤波器的最小平均P范数算法（VLMP）的收敛性能有待进一步改进的问题，提出了全解耦结构，即在二阶Voherra子系统满足最小分散系数（MD）准则的条件下使一阶子系统再满足MD准则的约束优化，优化过程周期滚动循环，最终实现了全解耦。仿真结果表明本算法收敛速度快，抗噪声能力强。

英文摘要：

The polynomial nature of second-order Voherra series yields poor performance in environments characterized by heavytailed distributions, because the quadratic terms amplify the damaging effects of present outliers and coupling between each order is very strong. It＇s necessary to improve the performance of the conventional Volterra filter least mean p norm （VLMP） algorithm. In order to solve this problem, this paper proposes a new fully decoupled structure which is a constrained optimization. In this process, the first-order Volterra subsystem meets the MD criterion on condition that the second-order Volterra subsystem meets the MD criterion, and this optimization procedure is roiling cyclical periodically, so the algorithm reaches the purpose the fully decoupling. Simulation studies show that the introduced method has fast convergence speed and is able to withstand strong impulsive noise.