Research on prediction model of tractor sound quality based on genetic algorithm

Abstract

The noise comfort of a tractor is a key factor in the evaluation of tractor comfort, and it has a significant impact on the evaluation of the overall performance evaluation of a tractor. This paper used the rating scale method to evaluate the sound quality of each operation state and position of the tractor, and obtain the evaluation value of the tractor noise. In addition, objective parameters of the sound (including the sound pressure level, A-weighted sound pressure level, loudness, sharpness, roughness, and fluctuation strength) were calculated. By considering objective parameters of sound quality as input and sound quality evaluation values as output, we established back propagation neural network (BPNN) model and support vector regression (SVR) model. Furthermore, the initial weights and thresholds of the BPNN model, penalty parameters, insensitive loss parameters, and kernel function parameters of the SVR model were optimized by genetic algorithm (GA). After verification of the experimental results, it was observed that the GA could improve the prediction accuracy of the model and significantly reduce the extreme errors. Compared with the prediction results of other prediction models, the GA–SVR model predicted the sound quality value of tractor noise with higher accuracy and stability.