Multiple fault diagnosis of analog circuit using quantum hopfield neural network

This paper address the multiple fault problem of analog circuit using quantum Hopfield neural network. The proposed quantum neural model, from the evolution of quantum states, gives a new interpretation of the associative memory mechanism in term of probability. The fault features are obtained by the wavelet packet analysis and energy calculation. The quantized ideal features of single fault and the actual features of multiple fault are regarded as quantum ground states and quantum excited states in the quantum space, respectively. Any excited state (multiple fault) in this space can be described as a superposition state of each quantum ground state with different probability amplitudes. The occurrence of this probability amplitude can be obtained by comparing the measurement matrix of the quantum-key-input mode with the measurement matrix of the quantum memory prototype. The numerical experiments offer a good explanation of the appearing probability of multiple faults.