Optimisation of Test Stimulus for Analogue Circuits Based on the AOA

Test stimulus is the key to the detection for analog circuits of complex integrated electronic devices, in order to improve the accuracy of analogue circuits detection, this paper proposes an optimization method for analogue circuit test excitation based on the Arithmetic Optimization Algorithm (AOA), which consists of the AOA and the Wilkes statistic as the optimization objective. Initially, the amplitude-frequency characteristics of the circuits are acquired via Pspice simulations. Subsequently, the AOA algorithm establishes vector individuals based on these amplitude-frequency profiles of the circuits. The Wilkes statistics are then employed to compute the intraclass and interclass distances across various circuit states, encompassing normal variations. The AOA, recognized for its heuristic nature, robust optimization capabilities, swift computational speed, and global optimization pursuit, is utilized to optimize the test frequencies. Once the optimal test frequency is determined via the AOA, it is employed as the input for testing the circuit under evaluation. The time-domain output signal of the tested circuit is subsequently classified using a linear support vector machine (SVM) to ascertain the efficacy of the proposed optimization method. To validate the efficacy of this proposed method, a four-op-amp dual-secondary bandpass filter circuit is chosen for experimental verification. The results obtained from the experiments showcase that the test excitation derived from this method achieves an accuracy of 97.02%. This accuracy surpasses that achieved by commonly used impulse sources when utilizing a simpler classifier like the linear SVM. The experimental outcomes robustly demonstrate the effectiveness of this method in optimizing test excitations for analogue circuits.