| Steady performance and transient quality are the two characteristics of vehicle motion dynamics. Transient quality plays an important role in guaranteeing the good market competitiveness of the products, which is usually defined as the undesired phenomenon aroused by improving performance or adding functionality, such as the poor vehicle stability, deviation from the route, bad road feel, inappropriate return ability, little prediction ability and weak consistency under different operating conditions etc. With the level of vehicle rising, it is inevitable to increase the design velocity and to improve the transient quality. It is the key difference between the advanced vehicle and the normal vehicle and also the main gap between the domestic band and the foreign enterprise.At present, due to the lack of theoretical basis as guidance, original equipment manufacturers have to rely on the real automotive subjective tests and tune to improve the transient quality of the vehicle, which depends on the accumulated experience of the engineers. It usually needs several years to form an excellent chassis platform after engineers upgrade the products many times, which negatively influences developing cost. Coming to the domestic band, as the lack of related experience accumulation and technical storage, it is not conducive to improve the products rapidly if the manufactures still use the traditional way to form the chassis development capability. The process of real car subjective tests and tuning can only be used in the last tuning period of automotive development. After the prototype vehicle was produced, some basic structure parameters have already been fixed, such as suspension hard-points, thus restricting the engineers to improve the quality of the products. Moreover, if there are some significant problems, the tuning work will not be carried through or the tuning results will not have enough market competitiveness, which increases the development risk. To sum up, engineers try to use computer simulation method to implement subjective evaluation and can control the subjective evaluation results at the early stage and during the whole design cycle of automotive. It is the urgent demand for chassis design and also has been an international hotspot and difficult issue for a long time.Aiming at the above requirements, scholars have conducted a lot of study on the correlation between the subjective evaluation and objective evaluation in the past. The research idea of this process is to implement subjective evaluation tests and objective evaluation tests respectively, after which the subjective evaluation results and objective evaluation indexes can be obtained. Then scholars try to get the correlativity between them and use the objective evaluation indexes which have high correlation to characterize the subjective evaluation results. In the process of automotive development, these indexes can be used as the design object for engineers.The thesis, aiming at the steering performance of automobile, proposes a new method to bring the subjective evaluation to the early stage of automotive development and control the subjective evaluation results, which contains the following directions, the first one is to find the characteristic indexes which can characterize the subjective evaluation results based on analysis of the intrinsic mechanism of steering subjective evaluation, the second one is to excite various kinds of subjective evaluation phenomenon based on the research of vehicle dynamics model oriented to subjective evaluation, the third one is to predict subjective evaluation score before prototype manufacture based on the model between subjective evaluation score and indexes. Specifically, the research of this thesis includes the following perspectives:To begin with, the thesis focuses on the mechanism of steering subjective evaluation of automotives. In the present literature, there is little systematic analysis of the intrinsic mechanism of subjective evaluation, and it is restricted to the objective evaluation criteria to implement tests and derive indexes which are used to characterize the subjective evaluation results. After communicating with the subjective evaluation engineers, we find that the subjective evaluation includes two aspects, one is the undesired response of the automotive, and the other is the deviation between the objective indexes and the preset characteristic curves. Different from previous studies, the thesis will study the typical evaluation phenomenon, operating scenario in the different evaluation system and related dynamic principle firstly, and then obtain the characteristic indexes which are used to characterize the subjective evaluation results based on the physical response derived from the subjective evaluation tests. The author will use coefficient of variation and correlation analysis to verify the indexes which represent the subjective evaluation phenomena, and finally refine the indexes system by computing the correlation coefficients between the different indexes.Secondly, the thesis studies the dynamics model oriented to automotive quality. Most of the existing simulation software and dynamics model adopt steady method to build the model. For example, the steering system model uses kinematics and mechanics method to compute the steered wheel angle and steering torque, and the tire model uses the relax length to the approximate carcass elasticity. These models, predicting the trend of the linear feature of the vehicle in the neighborhood of the steady state, are developed for the design of objective performance indexes, which cannot simulate the on center response and the ability to resist the interference of the automotive, so they cannot analyze the subjective evaluation of automotive, such as the straight driving evaluation. The thesis builds a dynamics steering system model which takes the Ackermann mechanism into account, a dynamics wheel model which takes the carcass elasticity into account and a suspension model which dynamically calculates roll center and accurately computes the position and pose of the spindle. This model, which can simulate the real response of the automotive more precisely, describes the steering friction and carcass elasticity carefully. With comparing the computing results with real car tests data and existing software, it is verified that the model not only can exactly simulate the field tests, but also can simulate lane keeping ability and stability of resisting small interferences.Thirdly, the thesis concentrates on studying the model between the subjective evaluation score and the characteristic indexes. It is necessary to build mathematical model between them to predict the subjective evaluation score at the early stage of automotive development and to balance and optimize the indexes with the predicted score as design object before the prototype is produced. The traditional linear regression analysis cannot describe the nonlinearity between subjective evaluation scores and objective indexes. Some scholars adopt nonlinearity function to compute accurate subjective scores based indexes, but most of them concentrated on just one phenomenon, which cannot compute the steering subjective score. The thesis, combine artificial neural network and fuzzy analytic hierarchy process together, builds a statistical correlation model between the phenomenon scores and the characteristic indexes and determines the relative weights of the different levels in the subjective evaluation system. As a result, the predicted value obtained, to a large extent, is consistent with that from the engineers.Fourthly, the thesis researches the improvement method of motion cue fidelity produced by the driving simulator. It is not allowed to install data measurement and collection machines during the real car tests to avoid influence the inherent characteristics of the automotive, so it is impossible to get the motion response when adopt this tests method. Meanwhile, the real car tests need lots of money and time, which are difficult to implement. Based on this, the thesis decides to use automotive driving simulator to finish the relative tests. To ensure the authenticity and credibility of the test, it is necessary to improve the fidelity of driving simulator. The author found there were few researches on the delay compensation of motion system after lots of survey, so the thesis proposes a new method to compensate the delay in motion system based on the state space prediction model, and further studies the modules of motion control algorithm. Consequently, the method and study solve the problem that the simulator cannot exactly and quickly implement the accelerated velocity.The thesis integrates the automotive dynamics model, characteristic indexes computing procedure and the model between the subjective evaluation scores and the characteristic indexes, successfully completes the subjective evaluation on the computer, and further validates the method proposed in this thesis. |
PDF Full Text Request