Characterization of a capacitorless current-mode bandpass biquad via an artificial neural network

An artificial neural network (ANN) is utilized to characterize a current-mode capacitorless bandpass biquad based on a simple single-stage CMOS OTA. The training data is automatically collected from HSPICE simulations based on the AMS´s 0.35 mu CMOS process. The trained ANN is able to indicate the feasibility of biquad´s requirements with low probability of type-I and type-II error. The ANN-generated circuit´s parameters, based on the feasible biquad´s requirements, are usually providing actual bandpass outputs with low percentage error, which is verified by H.SPICE-simulated responses.