以模糊理論實現一個具有學習功能的自動變速腳踏車系統

Abstract

自從18 世紀自行車被發明以來，經過兩百多年的演變，自行車不斷的被改 良。因應不同的用途，目前市面上充斥著各式各樣的自行車。而近年來更有許多 學術研究在探討自行車的自動變速，企圖設計出符合各種騎乘者需求的自動變速 自行車。

然而目前具有自動變速系統的自行車大部分都是已經先設定好規則和內建 參數，無法根據騎乘者的需求作調整，因此不適用於所有的騎乘者。本文除了採 用模糊理論制定一套換檔策略，透過微控器的訊號偵測與馬達的控制完成自動變 速以外，我們還提出一個學習方法，能夠依據不同學習者的騎乘行為調整參數， 以達到滿足不同騎乘者需求的目的。

The bicycles have been improved for more than 200 years since they were invented in the 18th century. Nowadays, there are different bicycles in the market for different objectives . There are many researches exploring in automatic shift control of the bicycles, intending for designing a bicycle that can shift gear automatically to meet different cyclists’s needs.

The bicycle currently with automatic shift control system almost have been setted rules and parameters already and cannot be modified according to cyclists’s needs, so they are not suitable for all the cyclists. According to the Fuzzy theorem, This paper develops a strategy that can shift gear automatically through signal sensing of the MCU and control of the motor. In addition, we present a learning method to modify the parameters according to the rider’s behavior such that meet different cyclists’s needs.