Instructions for Preparing Manuscripts

A major limitation of conventional microelectronics is the inability to perform pattern recognition at levels comparable with human perception. Software-based solutions of pattern recognition suffer from the delays required for transferring data between a memory and processor while hardware based solutions lack sufficient adaptability to continuously classify new patterns. Recently a new nanoscale circuit element called a memristor was physically realized having both memory storage and signal processing capabilities. The memristor allows data to be stored in multiple resistance states and provides a basis for the creation of circuitry to directly compare binary resistance data to binary voltage data. This paper reviews the state of the art in memristor development and proposes a memristor model and methodology for a pattern recognition circuit design.