Research On Optimal Design Of Ankle And Foot Orthosis Based On BP-GA Algorithm

The traditional product design method from the mapping process between the fuzzy user demand and the output object,to a large extent,is attributed to the semi empirical and semi formal thinking mode without system.This design process often has a "black box" fuzzy area,which seriously restricts the logic derivation and generation process of product visualization.For the development and design of ankle foot orthosis,a kind of human-machine rehabilitation aids,it also faces such a dilemma.Therefore,it is necessary to integrate a variety of existing design methods and technologies in different fields to better guide product design.Guided by the model design of ankle foot orthosis to meet the market expectation and user expectation,to avoid the fuzzy and uncertain problems in the design reasoning process,a shape optimization design of ankle foot orthosis based on the combination of BP neural network and genetic algorithm(GA)is constructed by combining the relevant knowledge in different fields such as perceptual engineering,statistical metrology,ergonomics and machine intelligent learning Method.First,the user’s preference perceptual image of ankle-foot orthosis is obtained through questionnaire survey and oral analysis,and the credibility of the questionnaire is analyzed.Based on Kansei Engineering’s theory of Kansei Engineering,the SD method was used to analyze it,combined with triangular fuzzy numbers and factor analysis to screen the dominant preference images;Then collect product samples through network collection,literature review,etc.to form an initial library of entity and concept samples;secondly,use morphological analysis to obtain the items and category characteristics of the components of the ankle and foot orthosis,based on this based on SPSS software Multi-dimensional scale analysis and cluster analysis to select superior samples;On this basis,the BP neural network is used to train the dominant preference images and dominant samples,the ankle-foot orthosis shape design elements are used as input layer parameters,and the perceived image evaluation values are used as output layer parameters,and the mapping relationship between the two is constructed;Taking the "hard" BP neural network as an example,the output value of the trained BP neural network is used as the fitness function of the GA(Genetic algorithm)algorithm to find the optimal product form combination in the sample solution space.The results show that on the one hand,this study provides an effective reference method for the design of ankle-foot orthosis oriented by user preference images,and on the other hand,it has certain reference value for the logical reasoning of the visual generation process of product design.