开关电流WTA神经网络

Switched Current Techniques Based Implementation of WTANN

本文研究了ＷＴＡ神经网络的开关电流技术实现。提出了ＷＴＡ神经网络的开关电流技术实现电路，分析了开关电流ＷＴＡ电路的动态行为。模拟实验和计算机仿真结果表明本文所提的开关电流ＷＴＡ神经网络是可行的。此开关电流神经网络可应用标准的数字ＶＬＳＩ工艺单片集成，且能在低电压如３．３Ｖ下工作。

The research on the implementation of winner take all neural network (WTANN) using switched current (SI) integrated circuit techniques is developed.The SI circuit architectures for the implementation of WTANN are presented.Meanwhile the analysis of the dynamic process of SI WTANN is also developed.The proposed SI circuit architectures are feasible illustrated by the experimental results and computer simulations.The proposed SI neural circuits are suitable for monolithic implementation with standard very large scale integrated (VLSI) digital circuit process,and can work at low power supply with 3.3 volt.