Application of an Adaptive-Network-Based Fuzzy Inference System for the Optimal Design of a Chinese Braille Display
學年 93
學期 2
出版(發表)日期 2005-02-01
作品名稱 Application of an Adaptive-Network-Based Fuzzy Inference System for the Optimal Design of a Chinese Braille Display
著者 Yeh, Fung-huei; Tsay, Huoy-shyi; Liang,Shih-hao
單位 淡江大學機械與機電工程學系
出版者 Singapore: World Scientific Publishing Co. Pte. Ltd.
著錄名稱、卷期、頁數 Biomedical Engineering: Applications, Basis and Communications 17(1), pp.50-60
摘要 This paper develops a forward-inverse prediction technique of an adaptive-network-based fuzzy inference system (ANFIS) to optimize the design of a relay-actuated Chinese Braille display (CBD) and to reduce the operation temperature. The key element in CBD, electromagnetic actuator assembled by relay, level, and plastic pin, must operate in lower temperature in order to have enough functional force exerted by relay to lift up Braille dots (pin-heads). However, electromagnetic actuators are often destroyed by thermal damage and the thrust force sensed from Braille dots is gradually reduced by the raising of operation temperature. Two design parameters that have the major influence on the thrust force and maximum operation temperature, including supply voltage and coil resistance of relay, were first analyzed by the thermal and electromagnetic numerical forward models to establish the basic databases for ANFIS. Using these databases, the optimal design of CBD can be forward-inversely predicted by ANFIS. As a verification of this forward-inverse mode, the forward predicted maximum operation temperature and the design thrust force inversely predicting the optimal supply voltage by ANFIS were compared with experimental data. The comparison indicates that the maximum operation temperature and the thrust force of both natural and force convection models achieved satisfactory accuracy. When two bell-shape membership functions were adopted, the best accuracy of the forward predicted maximum temperature of ANFIS reached is as high as 99.71% for natural convection model, and the best accuracy of the design thrust force for the inversely predicted supply voltage of ANFIS reached is also as high as 95.50% for force convection model. This investigation proves that the forward-inverse prediction mode developed for ANFIS with thermal and electromagnetic analyses can supply a useful soft computing technique for the optimal design of CBD.
語言 en
ISSN 1016-2372 1793-7132
期刊性質 國外
收錄於 EI
國別 SGP
出版型式 紙本

機構典藏連結 ( )