Macromodeling of the memristor using piecewise volterra series

The memristor, as the fourth basic passive circuit element, was first demonstrated using a TiO₂ layer sandwiched between two platinum electrodes forming a crossbar architecture by researchers at the Hewlett-Packard laboratories in 2008. The memristor is the nonlinear passive nano electronic device and holds great promise for a wide range of potential applications, because of its unique nonvolatile characteristic and memory effect. In this paper, the piecewise Volterra series macromodeling method was proposed for memristor, taking account of the nonlinear and memory effect. First several thresholds were set in the input signal amplitude according to the memristor characteristics, and the input signal was decomposed into several sub-signals using these thresholds. Then each sub-signal was processed separately by piecewise Volterra Kernels, and finally recombined together to produce the output. The experimental results and performance data demonstrate that the proposed method is efficient and precise.