Modelling of FinFET based ON bipolar cell of silicon retina

Abstract

—In this paper, we proposed a SOI FinFET based ON bipolar cell of silicon retina which plays an important role in neuromorphic applications like differential motion detection. Low power consumption along with biological plausibility is one of the crucial parameters of any neuromorphic retinal circuit as in the brain all biological systems consisting of billions of neurons consume power in terms of a few μwatts. First, the output of linkage cell between inner and outer plexiform layer i.e. bipolar cell is reconstructed and verified using developed ionic current based mathematical model which is established considering Hodgkin-Huxley model as the base model, and then biological behavior of the ON bipolar cell is mimicked through circuit using 180 nm, 90 nm and 45 nm Metal Oxide Semiconductor Field Effect Transistor (MOSFET). The power consumption in each case is obtained to be approximately 18.5 nW, 10.5 nW and 1 nW which is much larger than the power consumed by biological bipolar cell. Moreover, the same circuit is implemented with 32 nm Silicon on Insulator Fin Field Effect Transistor (SOI FinFET) device. It is observed that SOI FinFET based ON bipolar cell consumes 17.5 pW power which is less in comparison to MOSFET based bipolar cell. The output of bipolar cell is matched with biological nature of the waveform in each case.