Design of CNN cell with low-power variable-gm OTA and its application

In this paper, we propose a low-power variable-g_m operational transconductance amplifier (OTA) for cellular neural networks (CNN). The OTA use MOSFETS operating in the weak inversion region. By using a variable resistor, the transconductance g_m can be changed by the external bias voltage. The proposed OTA is verified by HSPICE simulations with the 0.18 mum standard CMOS device parameters. From the results, the power consumption is less than 7.0 muW with 1.2 V power supply. As an application of the proposed OTA, an oscillatory CNN circuit using two CNN cells with the proposed OTA is presented.