Research On Vehicle Interior Sound Quality Subjective And Objective Evaluation And Active Control Method

With the advance of modern industry, vehicle’s concept are developing from rideinstrument instead of walk to living space, so besides good performance such asenergy saving, safety and less pollusion, there is a growing emphasis on vehicle’s ridecomfort performance. Vehicle interior sound quality is an important index forintuitively reflecting the quality of automotive products and influencing customers’purchase psychology. Considering the variety of use demand of different types of cars,its sound quality characteristics are also quite different. Therefore, adopting effectmethod for evaluating and applying whatever active control strategies to improvevehicle interior sound quality according to the vehicle’s function and style areimportant research points. Though introducing advanced evaluation methods ofvehicle interior sound quality and active control technology, one-sidedness ofadopting the sound level A for measuring and describing sound is effectively avoided,and the disadvantage of passive noise control which has unsatisfied effect forlow-frequency noise control is well overcome, and many problems such as increasedproduction costs and longer development cycle lead by adding the absorption ofsound insulation material are perfectly solved, and individual customization of thevehicle interior noise based on the acoustic preference of driver and passengers couldbe commendably carried out as well.Combined with the national natural science foundation project called “Researchon Analysis Evaluation And Adaptive Active Control for Sound Quality of VehicleInterior Noise(50975119)”,evaluation and active control of vehicle interior soundquality are studied. In view to the related researches, the aims of this paper are tomaster the essence of interior sound quality spectral characteristics, investigate forscientific and reasonable evaluation and control methods, and solve the theory and engineering problems such as subjective and objective evaluation and the adaption ofactive control method for different type vehicles. Specific contents are as following:1) Though systemic analysis and study, objective and subjective evaluationindexes of vehicle interior sound quality are determined, then objective physicalacoustic parameters are analyzed and calculated, and the objective physical acousticparameters evaluation model which have consistent results with subjective evaluationhas been well established. Applying182interior noise samples collected by10familycars at different steady-state running conditions for the study, and according to therepresentatively selected50noise samples, objective evaluation with fourpsychoacoustic objective evaluation parameters, which contents loudness, sharpness,roughness and jitter degree based on semantic differential method have been done.Then respectively using multiple linear regression, BP neural networks, generalizedregression neural network and support vector regression principle, four kinds ofevaluation models for are vehicle interior sound quality are established. Predictionresults show that the effects of the4kinds of evaluation models.2) Considering the multidimensional characteristics, collecting the vehicleinterior noise evaluation indexes listed by the reference literature and automotiveenterprises in our nation and abroad, and combined with the describing habits ofdomestic people for sound feeling, many suitable evaluation indexes are selected byscoring and principal component analysis and the noise quality subjective evaluationscale which could accurately and objectively reflect human ear feelings areestablished. Then according to the validity analysis and truth test,12evaluationindexes which contents loud and quite, stiff and weak, motive, sharp and deep, fulland thin, irritability and calm, directivity, are adopt. Though subjective evaluationexperiments with the evaluation scale, difference analysis of evaluators gender, age,sound quality preference and whether acoustic experience are implemented, anddifference evaluation results caused by group identity are obtained.3) Combined with our national people’s acoustic preference, relationshipsbetween sound pressure amplitude of the24Bark domains and auditory perception are analyzed by subjective evaluation test and the relational model of12evaluationindexes and the sound pressure amplitude of the24Bark domains is established.UsingLMS and B&K pulse software,182vehicle interior noise samples is divided bycritical frequency band and then the amplitude of the sound pressure is calculated.Integrated with the subjective evaluation results of the12indexes predicted by theSVM noise quality evaluation models, relational model of12evaluation indexes andthe sound pressure amplitude of the24Bark domains are established by multiplelinear regression. And secondary sound source drive signal library of different soundquality characteristics is preliminary informed, then spectrum characteristic models ofluxury sound quality and sporty sound quality are established and relationshipsbetween sound audio domain information and subjective evaluation values areanalyzed.4) Comprehensive interior noise evaluation methods for different research needsare proposed by multi-target and multi-attribute decision method. Conventionalresearches on vehicle interior noise evaluation are limited to single-attribute ormulti-attribute of single stationary noise, and comprehensive researches onmulti-attribute of multi stationary noise is very rare. In this paper, series of method,such as fuzzy comprehensive evaluation, TOPSIS method based on entropy weight,multi-period multi-attribute decision-making methods based on linguistic informationand language multi-attribute decision-making method based on preferred program arereasonably applied and their effects are verified by many examples.5) According to the sound quality active the equalization control strategy of theluxury quality mode or sporty sound mode, different strength harmonic signals withsame phase or opposite phase of primary noise signals are transmitted based on theCenter frequency of the Bark threshold as dominant frequency, then the aim ofenhance or offset initial noise of each frequency could be achieved. Simulation modelof vehicle interior sound quality active control system is established and developed inMatlab/Simulink environment to verify the validity of the above algorithm. And basedon the introduced theory, vehicle interior sound quality active controller is developed by using80C51F120type microcontroller hardware. The active control test results forthe noise around two ears of copilot of a domestic car in4kinds of steady statedriving conditions show the perfect validity and feasibility of the sound quality activecontrol algorithm.