A Neuro-Intelligent Technique for Circuital Analysis of Analogue Phase Detector

The integration of artificial intelligent techniques with modern circuitry is fast becoming an attractive technique. This study, therefore, develops a Resilient Backpropagation Algorithm (RBPA) based artificial neural network (ANN) to analyze and predict the behavior of an analogue phase detector circuit. RBPA efficiently addresses issues posed by the magnitudes of partial derivatives in traditional backpropagation algorithm. The analogue phase detector, designed using Proteus software, provides the training data with which the network was trained and the performance evaluated against the performance obtained under three other popular algorithms under same network configurations. The Resilient Backpropagation Algorithm (RBPA) based artificial neural network outperformed the ANNs trained with the other algorithms as it attained a training Variance Accounted For (VAF) of 98.64 and a testing VAF of 96.26. The results demonstrate that RBPA significantly enhances the learning and generalization capabilities of ANNs, making it a robust and reliable method for circuit analysis and prediction, a technique which helps to eliminate laborious repetition of computer-based simulation.