Kansei Evaluation Model of Tractor Shape Design Based on GA-BP Neural Network

To mine users’ perceptual demand for product shape, it is very important to build the relational model between design elements of product shape and users’ Kansei evaluation. We apply Kansei engineering to the shape design of wheeled tractor. The design elements obtained by morphological analysis constitute the input layer, and perceptual semantic evaluation obtained by semantic differential method constitutes the output layer. Genetic algorithm-back propagation (GA-BP) neural network is used to construct the relational model between design elements of product shape and users’ Kansei evaluation. Experiment shows that the predicted values using the training samples are consistent with Kansei evaluation values based on GA-BP neural network. Meanwhile, the relative error between the predicted and measured values of users’ Kansei evaluation using the testing samples is less than 3%. However, There is a larger deviation between the predicted values using the training samples and Kansei evaluation values based on BP model. Furthermore, the relative error between the predicted and measured values of users’ Kansei evaluation using the testing samples is more than 10%. Comparison GA-BP with BP neural network modeling shows the perceptual evaluation model based on GA-BP is superior to BP network model. Therefore, the model is capable of accurate prediction of users’ Kansei evaluation about product shape utilizing GA-BP neural network modeling and can be used to guide product shape design. This will not only improve the efficiency of product research and development, but also enhance enterprises’ competitiveness.