Time division digital programmable OTA for cellular neural networks

The design of a digital programmable transconductance amplifier (TD-DPTA) for cellular neural networks is shown in the paper. The proposed approach is based on a time division strategy with the main goal of a reduction of the area by the minimization of the number of the analogue multipliers inside each cell. So, the new CNN cell is composed of a single TD-DPTA only, which is able to feed the whole neighborhood by the proper weighted current contribution by using a time division approach. The paper presents the design of a new DPTA architecture as well as a multiplexer circuit tailored for this particular application.