A hybrid fuzzy neural decoder for convolutional codes
Meng Wu, Wei‐Ping Zhu, S. Nakamura · 2002
In this paper we propose a hybrid fuzzy neural network for decoding of convolutional codes. The decoding process will be completed by classifying the proposed network instead of conventional decoding methods such as Viterbi algorithm. According to the encoding principle of convolutional codes, the size of the network is determined dynamically through clustering. Logic operations are also used in the network, so the training speed of the network is very fast: only one or several iterations are required. Moreover we define fuzzy membership function for each hidden node, which enhances the associative capability of the network, thus improving the rectifying capability. For the small constraint length we study the performance of the proposed method and compare it with the Viterbi algorithm.