本論文的目的是發展一個以數位訊號處理器為基礎之類神經網路 控制模組化輕型電動車之六相永磁同步馬達驅動系統。首先,介紹模 組化輕型電動車的概念,再來介紹包含不確定項之輕型電動車和六相 永磁同步馬達的動態。接著利用一個32位元的定點運算數位訊號處理 器(TMS320F2812)來實現本論文所提出的控制器及控制一六相馬達驅 動器。由於包含滾動阻力、空氣阻力、參數變化、外力干擾和摩擦力 的負載轉矩會影響控制的準確性,因此本論文提出具有線上學習能力 以及良好的強健性的Elman類神經網路控制器與機率模糊類神經網路 控制器兩種智慧型控制器,來達到輕型電動車應用所需求高的控制效 能。最後,根據實驗結果來驗證本論文提出的模組化輕型電動車之六 相永磁同步馬達驅動系統之性能。The purpose of this thesis is to develop a digital signal processor (DSP)-based neural network controlled in-wheel motor system using a six-phase permanent magnet synchronous motor (PMSM) drive for modularized light electric vehicle (LEV). First, the concept of modularized LEV is introduced. Then, the dynamics of modularized LEV and six-phase PMSM drive system with a lumped uncertainty are briefly introduced. Next, a 32-bit fix-point DSP, TMS320F2812, is adopted for the implementation of the proposed control system and to control a six-phase PMSM drive. Moreover, the control accuracy is much influenced by the load torque form the rolling resistance, the air drag, the parameter variations, the external disturbances, and the friction force. Therefore, two intelligent controllers, Elman neural network control (ENN) and probabilistic fuzzy neural network (PFNN) controller with on-line learning capability and robust control characteristics, are proposed to achieve the required high control performance of LEV. Finally, some experimental results are illustrated to show the validity of the proposed control schemes of six-phase PMSM drive system for modularized LEV.