Abstract
This paper presents a study on hybrid linear-fuzzy model, based on the techniqueof fuzzy reasoning and mathematical quantification theory No.1. This model is for the subjective evaluation model of drivers of vehicles under different driving conditions. The linear equation generated by mathematical quantificationis for the linear factors of drivers' subjective evaluation. The other model, fuzzy reasoning model deals with its nonlinear factor. In order to easily realize this model into a simple system which does not require any specific exclusive fuzzy hardware like a fuzzy chip, the technique of simple forming fuzzy variables, namely fuzzified labels, and membership functions has been proposed. In generating fuzzy reasoning knowledge, average values and variant values of observed data are considered. The advantage of the hybrid model is best seen when the linear predictor cannot estimate nonlinear parameters in the simple modelling method. An example involving a vehicle-driversystem is given to show the merit of this modelling technique.
