【作者】 张金焕；

【导师】 刘泉；

【作者基本信息】 武汉理工大学 ， 通信与信息系统， 2006， 硕士

【摘要】 随着大规模集成电路技术的不断发展,新一代微控制器在各方面性能上都有长足的进步,嵌入式处理器更加适合数控领域的开发。其稳定性高,没有对资源的过多占用,没有其它不相关硬件的干扰,从而使系统很纯粹地用于所需的加工控制;嵌入式微处理器的集成度高,片上集成多种功能模块,网络支持进一步完善,集成了更多种类的网络接口;微控制器和处理器开发手段和调试平台越来越完善,嵌入式实时操作系统的广泛应用,对于微控制系统的开发提供了良好的任务管理平台和底层驱动平台,这为上层软件模块的可靠简便开发和管理提供了有力的保证。 而在其伺服控制算法方面,PID(比例、积分、微分)控制是在工业过程控制发展史上最早发展起来的控制策略之一,由于其算法简单直观、鲁棒性好和可靠性强,被广泛应用于工业过程控制,尤其适用于可建立精确数学模型的确定性控制系统,然而PID控制的参数整定是一个十分繁琐费时的工作。计算机技术和智能理论的发展为复杂动态不确定系统的控制提供了新的途径。 本文研究在嵌入式系统理论和柔性嵌入式数控系统理论的基础上,提出了基于嵌入式、开放性、网络化的数控体系结构,并给出了在硬件和软件方面上的体系结构的设计,尤其是对ARM部分及其中的G代码解释模块的设计作了较为详细的介绍。以ARM及DSP构建系统的主控平台,通过模块化的软硬件结构设计思路,在ucos-Ⅱ的支持下,克服了当前数控平台的缺陷,实现了系统的实时调度,提高了系统的可靠性及稳定性,实现了系统的多种网络支持,符合数控系统模块化、柔性化和网络化的发展需要。 本文还对智能理论(模糊理论和神经网络)做了分析和研究,并在传统PID控制算法的基础上,使其用到嵌入式数控系统的伺服控制部分PID参数的优化方面。结合智能理论和方法设计智能PID和进行PID的智能整定,来解决常规PID对实际工业生产过程往往具有非线性、时变、不确定性、难以建立精确数学模型的对象不能达到理想控制效果的问题。对利用模糊控制和神经网络理论与PID结合来实现其参数的智能整定进行了研究和分析,并借助于MAFLAB/SIMULINK对选取的控制对象进行了仿直和分析。更多还原

【Abstract】 With the continuing development of Large Scale Integration technology, the new MCU (Micro Computer Control Unit) has great progress in every performance and embedded processor is more suitable to development of digital control. It has great stability, and it does not need excessive resource and is not disturbed by relate hardware. So the system meets needs of process control. The embedded MCU is highly integrated and there are many kinds of functions on the chip. The network function has been improved and also there are many kinds of interfaces integrated. The developing means and testing plat form are more and more perfect. The widely use of RTOS (Real Time Operating System) provides MCU development with good platform of task management and bottom drive, and also guarantees the steady and convenient development and management of apply software.In the aspect of servo control algorithm, PID (proportion, integral, differential) control is the earliest strategy developed in industry process control. It is widely used in industry process control, especially in the determinant system with exact mathematical model for it is simple, robust and steady. The tuning of PID parameters is a very long and taking time process. The development of computer technology and intelligent theory provides new methods to control complicated and dynamic systems.This paper brings forward an embedded, opening and latticing numerical control structure based on the theory of Embedded System and Flexible Embedded Digital Control System and describes the structure design in hardware and software, especially introduces G code analysis module of ARM in more detail. The control platform based on ARM and DSP could overcome defects of current digital control system and realize real - time reservoir operation by modularize design of hardware and software and support of ucos-II. Also it improves the reliability and stability and it could have many kinds of network support, which is accord with the development of modularization, flexibility and latticing.Also this paper dose analysis and research on intelligent theory including Fuzzy theory and Neural Network, which is used in the aspect of PID parameters optimize of servo control based on the conventional PID. In order to solve the problems of conventional PID that it has non-linearity, time variation, uncertainty and is difficult to constitute exact math model in practical industry manufacture, this paper mainly combines Fuzzy Control and Neural Network with PID to tune parameters. Also it uses MATLAB/SIMULINK to do simulation and consequent analyses about selected control object.更多还原