提出了一种隶属度可编程的模糊识别器。该识别器采用“求和－排序”模糊操作进行模式识别。其中，隶属函数可以根据需要进行编程，以满足不同应用场合的要求。采用单层金属、单层多晶的２μｍＮ阱标准数字ＣＭＯＳ工艺成功地制作了该模糊处理器中的核心单元电路的芯片。测试结果表明，这些电路芯片的性能很好，完全可以满足处理器的性能要求 A fuzzy recognizer with programmable membership is proposed. The recognizer uses “sum-and-order” fuzzy operations for pattern recognition. Among them, membership functions can be programmed as needed to meet the requirements of different applications. The chip of the core unit circuit in the fuzzy processor was successfully fabricated using a single-layer metal, single-layer polycrystalline 2μm N-well standard digital CMOS process. The test results show that the performance of these circuit chips is very good, fully meet the performance requirements of the processor