Implementation of neural network function to neuron-type conducting polymer and its characteristics

具有神经元样结构的导电聚合物已经聚合，可以连接成网络。如果在导电聚合物的网络中增加三个功能：I、多层、II、非线性V-I特性(阈值函数)、III、通过在网络路径中传递的信号来控制重量(即导电性)，则可以制造类神经元器件或神经网络器件。在本研究中，研究了添加三个功能的方法。AND和OR逻辑电路由单个神经元状导电聚合物建立。并利用纠错学习方法建立了多层网络的逻辑电路。用扩散法和交叉电位法对突触类导电聚合物作为突触功能器件进行了研究。作为神经递质的掺杂剂在突触间隙中通过扩散和迁移被注入和传递。

The conducting polymers with neuron-like pattern have been polymerized and can be connected to make a network. The neuron like device or neural network device can be fabricated if three functions are added to the network of conducting polymer: I. Multilayer, II. Non-linear V-I characteristics(threshold function)), III. Weight (i.e. conductivity) control by using signals that pass in the paths of the network. In this research, the methods to add three functions have been investigated. AND and OR logical circuits have set up by a single neuron-like conducting polymer. Furthermore AND logical circuit of multilayer network has been built up by the error correcting learning method.. Synapse-type conducting polymer has been estimated as a synapse-function device using diffusion method and cross potential method. Dopant as Neurotransmitter has been injected and transmitted by diffusion and migration during a gap as synaptic cleft.