The primary construction and its implementation of multidimensional asynchronous pattern classifier were presented. On line optimization of classification threshold can be finished by a learning program based on Bayes minimum error law or Bayes minimum risk law. 介绍了异步多维模式分类器的基本结构和实现方法,根据Bayes最小错误分类原则或Bayes最小风险分类原则,可通过学习程序实现分类阈值的在线寻优。
The built-in intelligence can be realized by the hardware and software of a control system, which includes multidimensional synchronization pattern classification strategy, unit control rule system, learning algorithm for the compensation to the dissonance and time-variation of parameters of photoelectric sensors, precondition learning program. 内环智能体系由系统硬件实现,包括多维异步模式分类策略、单元控制规则系统、对元件性能参数的离散性和时变性的补偿策略及算法、机器现场的训练学习程序。