Title :
Determining optimal switching speed for memristors in neuromorphic system
Author :
Yakopcic, C. ; Taha, T.M.
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Dayton, Dayton, OH, USA
Abstract :
As non-volatile memory based on resistive switching becomes more mature, memristor devices with very fast switching times are becoming more prominent. However, this reported work shows that memristor devices with slow switching times (of about 10 μs) are more appropriate for use in neuromorphic systems. This is done by modelling a series of memristors that differ in their switching time. Simulation of a memristor-based neuromorphic circuit is performed using each of these modelled devices. Devices with a high switching speed require unrealistically small voltage pulses to incrementally change the memristor resistance.
Keywords :
memristor circuits; neural nets; memristor based neuromorphic circuit; memristor resistance; neuromorphic system; optimal switching speed; switching time;
Journal_Title :
Electronics Letters
DOI :
10.1049/el.2015.1145