Nonlinear Solution for Digital Instrument Interface Based on Kansei Engineering

The digital instrument interface designers in industrial circle must face the typical nonlinear and uncertainty of the relationship between Kansei images (KIs) and product attributes (PAs), and the shortage of the statistical sample. It is conducted thorough research to the problem of nonlinear processing method, pointed out a new way that using combinatorial methods to solve the complex problem of small samples and poor information, and then put forward the grey-neural network (GNN) to deal with the relationship between KIs and PAs for digital instrument interface, to determine the best design combination of interface design elements for matching a given Kansei image represented by a word pair. A case study involving the design of digital tachometer interface is presented to demonstrate the proposed method. It was shown a strong feasibility through the test comparison with the grey and BP neural network prediction.