Loudspeaker fault detection using artificial neural network

Traditionally, loudspeaker´s quality control has been done manually and inspection of loudspeaker faults is time consuming and causes error in the quality evaluation. In order to reduce the time consumption and errors in the quality evaluation, in this research work, a simple loudspeaker diagnosing system is developed based on the harmonic distortion. The faulty and normal loudspeakers are tested using the sound emanated from the loudspeaker in the frequency range between 20 Hz and 20,000 Hz. A fast Fourier transform (FFT) is applied on the recorded signal to transform from time domain to frequency domain and the frequency spectrum is obtained. From the frequency spectrum, the total energy in the first 6 frequency bands are computed and chosen for further analysis. These frequency band energy signals obtained are then used as features for training the neural network. A simple neural network model is developed for the automatic detection of loudspeaker faults. From the result it is observed that the proposed method is able to classify the faults with an accuracy level of 82%.