一种改进的扩展窗喷泉码不等差错保护编码算法

针对传统不等差错保护（Unequal Error Protection,UEP）算法在加性高斯白噪声信道中存在误码平台的问题，提出了一种改进的扩展窗喷泉码（Expanding Window Fountain,EWF）编码算法。通过为不同重要等级数据加窗的方式，使得重要等级较高的信息节点能够以更高的概率连接至更多的校验节点。在编码过程中分开记录不同窗内信息节点的度数值，并优化校验节点选取信息节点的方式，使得度数值相对较小的信息节点始终能被优先选取，从而提高了每个窗内信息节点的平均度数值，为各重要等级的数据提供了额外的增益。仿真结果表明，改进的EWF算法可将不同等级数据的误码平台降低1~2个数量级，且数据的等级越高，保护性能越强。

An improved expanding window fountain codes encoding scheme for unequal error protection

Traditional unequal error protection(UEP) schemes experience error floor in additive white Gaussian noise channels,so an improved expanding window fountain(EWF) encoding scheme is proposed.By windowing the data according to their protection requirements,the more important part of information nodes can be connected to more check nodes with high probability.The degrees of information nodes in different windows are recorded separately in the encoding process,and the selection method of check nodes is optimized.Then each check node will preferentially select information node as its neighbors from small degree sets.Thus the average degree of information nodes in each window is increased and additional gain is obtained for data of different important levels.Simulation results show that the proposed EWF scheme can decrease the error floor by 1~2 orders.Furthermore,the higher the data level is,the better the bit error rate performance can be achieved.