Modeling and Characterization of Resistor Elements for Neuromorphic Systems

Physical structures changing their resistance in operation can serve as a basis for making elements of neural networks (synapses, neurons, etc.). The processes inducing changes of resistance are rather complicated and cannot be described readily. To demonstrate the potential of this sort of variable resistors it is possible to substitute a complex physical system by a simple mathematical model reproducing the important behavioral characteristics of the actual system. A simple resistor element whose state is defined by a single scalar variable is taken as a model unit. Equations responsible for changes of the state variable are determined. Different functions and parameters that can enter these equations are discussed. Combinations of such elements and conventional electronic components are considered. Measurement methods for variable resistors are investigated. Experimental data are used to determine characteristics of a particular type of variable resistor, metal-insulator-metal structures with amorphous titanium dioxide as insulator. Specific sets of functions defining the “voltage-current” experiment-resembling behavior of a resistor element are presented.