OPTIMAL DESIGN OF CMOS BIQUAD BANDPASS FILTER USING MLP-ANN

In recent days the requirement of various kinds of filters has been increased for wirelesscommunication and the primary effort on a resourceful practice to design and synthesize the Bandpass filter for different wireless application is presented. This paper proposes the 2nd order CMOS based active bandpass filter with biquad structure and characterized by Artificial Neural Network (ANN). The cross-coupled PMOS transistors are utilized to achieve bandpass response and the centre frequency of the designed filter is 28 KHz. The proposed structure improves the filter performance by 1.2 V and produces 1.62 uV/sqrt (Hz) while consuming 47.19 nW. At the same time the inductorless approach makes the circuit non bulkier for achieving bandpass response for variety of applications in communication. The design parameters of bandpass filter is optimized by Multi layer perceptron Neural Network with the parameters obtained from simulation results. However both the results are merely equal and hence indcates that our design is optimal.