【作者】 陶亮；

【导师】 张南纶；

【作者基本信息】 武汉理工大学 ， 控制理论与控制工程， 2005， 硕士

【摘要】 医疗诊断推理专家系统是当前人工智能研究中最活跃的分支之一,它实现了人工智能从理论研究走向实际应用,从一般思维方法探讨转入专门知识运用的重大突破。医疗诊断逻辑推理机作为医疗诊断推理专家系统最核心的部分,其性能的好坏决定该专家系统的成败。主要的研究成果如下: 采用泛布尔代数公理体系作为理论依据研究医疗诊断逻辑推理机的原理,提出了如何运用逻辑非运算来实现推理过程的简化,并介绍实际的医疗诊断逻辑推理机的设计方案与方法。 开展了医疗诊断逻辑推理机推理策略的研究。讨论了基于初步诊断与鉴别诊断的正反推理过程与基于病症-疾病映射关系的泛布尔函数式的匹配比较的研究。同时研究了逻辑非运算的适用范围与使用方法,以及如何采用逻辑非运算来实现泛布尔函数式的简化。 整个推理机由上位机人机交互界面和下位机硬件推理机两部分组成。前者的主要功能是为用户手动输入疾病症状提供一个交互式平台,可以完成向用户提问,并将用户回答的结果简单处理后传给下位机。后者的主要功能是完成推理,并将推理的结果回传给上位机显示。上位机人机交互界面采用Visual Basic编程,下位机硬件系统采用“ARM单片机PLC2124+XCR3128型CPLD+USB接口芯片PDIUSBD12”的构架模式。 完成医疗诊断推理机推理比较部分的软件设计。基于ISE Foundation 6.1软件平台,采用VHDL硬件描述语言对CPLD硬件推理机进行了研究和设计。基于ADS集成开发环境,采用C语言实现了PLC2124的系统流程控制、数据通讯和协助CPLD完成推理。 完成医疗诊断推理机USB通讯软件编程。在基于Windows DDK的USB硬件驱动平台下,对推理机下位机硬件与PC互联进行了研究和设计。针对软件的模块化,着重实现了USB设备固件程序、USB设备驱动程序与USB设备通讯应用程序三个子模块程序的编写。基于Visual Studio 6.0软件,设计了上位机人机交互界面,并研究了上位机人机交互界面与下位机硬件设备间USB通讯方式;制定了符合本系统的通讯协议,并在该通讯协议的基础上,实现了上、下位机间的数据交换。更多还原

【Abstract】 Medical Diagnosis Inference Expert System is one of the active embranchment of artificial intelligent research at present. It is an importance breakthrough that the Expert System realized practice application of artificial intelligent from academic research and realized expert knowledge application from common thinking way discussion. As hardcore of Medical Diagnosis Inference Expert System, stand or fall of Medical Diagnosis Logic Inference’s capability determined the Expert System. Main research achievement is as bellows:Medical Diagnosis Logic Inference’s theory is researched based on Pan-Boolean algebraic on Axiomatic, how to realize consequence applying Non-logic is put forward, and an practice design scheme and method of Medical Diagnosis Logic Inference are introduced.Research on conducted strategy of Medical Diagnosis Logic Inference is developed. We discussed forward reasoning and backward reasoning based on basic and differential diagnosis, and studied matching compare of Pan-Boolean function based on symptom and disease mapping. Simultaneity, applicable range and using way of Non-logic are studied, and how to simplified Pan-Boolean Function adopting Non-logic.The integrated Inference makes up of two parts: the host human-computer interaction interface and hardware Inference. The former, which accomplish enquiries for user and transmit answer’s result of user by simply dealing to MCU, provides an alternating platform for user inputting disease’s symptom manually. The later completes conduction, as well as return conducing result to the host. The host human-computer interaction interface adopt Visual Basic, while the hardware system apply PLC2124 of ARM MCU, XCR3128of CPLD and PDIUSBD12 of USB interface chip.The soft design on conducing and comparing of Medical Diagnosis Logic Inference is completed. Based on ISE Foundation 6.1 soft platform, we made research and design on the CPLD hardware Inference applying VHDL Hardware Description Language. It realized system flowing control of PLC2124 communication by applying C language, and accomplished conduction by assisting CPLD, based on ADS integrated exploration environment.The USB communication program of Medical Diagnosis Logic Inference is completed. In the USB hardware drive platform based on Windows DDK, mutual link between the hardware Inference with the host is researched and designed. Aiming to the soft modularization, we mainly realized three blockettes: the firmware program of USB device, the drive program of USB device and the communication application program of USB device. Based on the Visual Studio 6.0 software, we designed the host human-computer interaction interface, and studied the USB communication way between the host human-computer interaction interface with the hardware; the communication agreement fitting for the system is established, and realized data transmission between the host with the hardware.更多还原