研究了车辆自动变速器的换挡建模问题 ,介绍了利用径向基神经网络 (RBFN)的换挡建模方法和算法 ;为了满足建模过程的在线运行 ,设计了 DSP处理器 TMS32 0 C31和 80 C196双 CPU硬件系统 ;测试结果表明 :RBFN网模型能够很好从驾驶员的换挡数据中获得换挡规律 ,利用 DSP实现 RBFN算法满足工程要求 The gear shift modeling of vehicle automatic transmission is studied. The gear shift modeling method and algorithm using Radial Basis Function Neural Network (RBFN) are introduced. In order to meet the online operation of the modeling process, DSP processor TMS32 0 C31 and 80 C196 double CPU hardware system. The test results show that the RBFN model can get the shift schedule well from the driver shift data, and the RBFN algorithm can meet the project requirements