LM-SVM-DT Based Working State Recognition for Washing Machine's Audio Signal

• In order to make a reasonable and effective judgment on the quality inspection and fault diagnosis of intelligent electrical appliances, this paper proposes a method of working state recognition for washing machine based on audio signal, named as LM-SVM-DT. The whole working process of the washing machine is divided into four basic states: water intake, soaking, washing and dehydration. The Log-mel features of the audio signal after bandpass filtering are extracted and modeled by the decision tree classification method based on SVM. That is, soaking and non-soaking states are separated at first, then washing and non-washing states are separated in non-soaking states, and finally water intake and dehydration states are separated in non-washing states. Taking the standard-washing-mode of a certain type of washing machine as an example to verify the algorithm, the experimental results show that the state recognition rate is as high as 0.9920. The results show that the model proposed in this paper is effective and feasible.

Subject headings

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Signalbehandling (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Signal Processing (hsv//eng)

Keyword

audio signal

decision tree

Log-mel

SVM

Working state recognition

Publication and Content Type

ref (subject category)

kon (subject category)