摘要
本文提出了一种将概率神经网络进行改进井使其用于信源一信道联合编码的方法。改进后的网络可以利用信源中的相关性，对噪声有更好的鲁棒性。它的复杂度低，结构设计简单，并可以灵活地改变以适应不同的信源模型。
An improved probability network is proposed and used in joint source-channel coding. Using the Correlation in source, the improved network is more robust to the noise. Its complexity is low and the construction is simple. Furthermore, it can be altered easily to adapt to different kinds of the source models.
出处
《通讯和计算机（中英文版）》
2005年第6期68-70,共3页
Journal of Communication and Computer
