关键词: 换挡品质, 径向基函数神经网络, 评价系统, 评价指标, 双离合器式自动变速器

Abstract: At present，subjective evaluation is mostly adopted for shifting quality evaluation, so it is easily influenced by the testing personnel, and the effects of the shifting process on other performances are hardly considered. Based on the vehicle of dual clutch transmission (DCT), with the comprehensive consideration of multiple factors including dynamic performance, economy, durability of transmission system and comfort, combining with shifting time, acceleration, jerk, energy density, etc, the shifting quality evaluation indexes are put forward. The relationship between subjective and objective evaluations of comfort is established through radial basis function (RBF) neural network, and the influences of different parameters on the network performance are examined, then the trend between training error and number of neurons and the difference between original data and training data are presented. Based on this, the shifting quality evaluation system is developed by using programming interface. The shifting process is simulated by combining with a driver’s intention. The results show that the system developed can effectively eliminate the influence of subjective factor and increase the objectivity of shifting quality evaluation.

Key words: Dual clutch transmission, Evaluation index, Evaluation system, Radial basis function neural network, Shifting quality