Abstract: The domestic auto seat noise detecting is mainly in anechoic chamber in which require the environment of noise does not exceed 30 dB.However,this method of defect is building an echoic chamber such a high cost,low detection efficiency,operator skill affect measurement accuracy.In view of the above defects,this paper proposes a stochastic resonance based on singular value decomposition of on-line detection method of automotive noise signal,can filter out outside noise interference,effectively extract the seat vibration noise characteristic signal,for subsequent online assessment laid a solid foundation.Simulation and example results show that the singular value decomposition of stochastic resonance is denoising method than using singular value decomposition and more stable stochastic resonance method detects the signal frequency is more accurate,can enhance the amplitude of the signal,to better detect the weak signal submerged by noise.

Key words: auto seat noise, Singular Value Decomposition(SVD), stochastic resonance, detection system