Research On Evaluation Method Of Static Comfort Of Automobile Seat Based On Fuzzy Neural Network

With the continuous release of national macroeconomic policies,the national economy continued to maintain a good momentum of development.Under the influence of this,China's auto industry to increase the intensity of reform,product structure adjustment and update the pace continues to accelerate,production and sales growth year by year,and China is still in the car popularity period,long-term demand is still large space.The competition between the OEMs prices has been carried out at various levels,including vehicle modeling,configuration,price,after-sales service,the past simple price war effect has been greatly reduced,and competition to a higher level,including ergonomics,active safety,comfort and so on.In the driving process,the car seat and the human body contact area of the largest contact time is the longest,whether it is good or bad directly affect the occupants of the safety and comfort.This paper mainly discusses the static comfort evaluation method of automobile seat,the main contents are as follows:(1)Improve the test process;through the BMI scale combined with the height of the selection of the appropriate test person,the initial stage of the trial to each of the subjects to establish a scientific definition of comfort,design based on the subject of the subjective scale,through the real vehicle simulation platform Test the different subjects by 10 different seat body pressure distribution,access to hundreds of sets of test data for follow-up modeling.(2)Optimize the subjective evaluation.Through the unified treatment of subjective scoring,the scoring of the subjects is at a relatively rational level,weakening the effect of the subjective factors and improving the prediction accuracy of the final model.(3)Local feature extraction;through the actual test research,to understand the short-term static ride experience in the more sensitive parts of the hip area and thigh region,the use of MATLAB function tool to extract the pressure characteristics of the two regions,and clear its physical meaning,To ensure the generalization of the evaluation model.(4)Based on the prediction results of the algorithm selection;in a variety of algorithms to choose from the case,try to get the results in the actual training model to select the appropriate combination of algorithms.(5)Based on ANFIS evaluation system modeling;try to use ANFIS editor and script program combined training,in order to shorten the training time,improve training efficiency.The average error rate of the prediction model based on the fuzzy neural network is 7.35% for the 10 sets of test samples.When the results are consistent with the subjective evaluation results,the short-term seat static comfort rating is realized,The side verifies the feasibility of improving the comfort of the whole chair by modifying the local seat feature.At the end of the paper,the future research on seat comfort is prospected.The development of computing ability,depth learning and intelligent algorithm will greatly promote the research process of seat comfort.