长约束度卷积码译码器神经网络结构

A Neural Network Architecture for the Decoding of Long Constraint Length Convolutional Codes

给出的神经网络译码器是为长约束度（Ｋ≥１１）卷积码译码而设计的。Ｖｉｔｅｒｂｉ译码和序列译码是两种最大似然译码方法，虽然这两种技术能有效地提高误比特率性能，但它们都存在局限性。另外，只要在神经元和数字异或门单元之间建立局部连接，就能非常容易地直接用超大规模集成电路（ＶＬＳＩ）实现硬件。

The neural network decoder presented in this paper is designed for long constraint length convolutional codes (K≥11). The two most effective methods of decoding convolutionally- encoded data are Viterbi decoding and sequential decoding - two maximum likelihood probabilistic decoding techniques. Both techniques are very effective at improving the Bit Error Ratio performance over un - encoded transmission of data, but each method has its limitations. Moreover, With only local connections between neurons and digital Ex - OR cells, direct hardware implementation in a VLSI ASIC is feasible.