• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2011, Vol. 33 ›› Issue (4): 154-158.

• 论文 • 上一篇    下一篇

基于DSP的BP神经网络PID控制器的设计

刘〓畅   

  1. (华南理工大学机械与汽车工程学院,广东 广州 510640)
  • 收稿日期:2010-04-25 修回日期:2010-08-07 出版日期:2011-04-25 发布日期:2011-04-25
  • 作者简介:刘畅(1986),男,江西瑞金人,硕士,CCF会员(E200015631G),研究方向为高性能嵌入式控制系统。
  • 基金资助:

    国家自然科学基金资助项目(50875090);国家863计划资助项目(2009AA04Z111);国家科技支撑计划资助项目(2007BAH13B06)

Design of BP Neural NetworkPID Controllers Based on DSP

LIU Chang   

  1. (School of Mechanical and  Automotive Engineering,South China University of Technology,Guangzhou 510640,China)
  • Received:2010-04-25 Revised:2010-08-07 Online:2011-04-25 Published:2011-04-25

摘要:

针对BP神经网络PID控制算法的复杂性及实现的困难性,本文提出了一种使用DSP芯片来实现的方案,外围功能接口则由辅助芯片FPGA来完成。利用TI公司提供的RTOS(DSP/BIOS)快速开发出该控制器原型,并通过对伺服电机的转速控制实验,对比传统的PID控制后,证明了该方案的实时性及控制性能都能满足工程需求。

关键词: BP神经网络, PID控制器, DSP, 电机控制

Abstract:

As for the problem that BP neural network PID control algorithm is complex and hard to implement, a scheme that uses DSP to implement the algorithm is presented, and the external function interfaces are implemented in FPGA. The controller prototype is quickly designed with the help of RTOS(DSP/BIOS) provided by TI. Compared with the traditional PID controlling, the experiment  using this controller to control the velocity of the servo motor shows that the scheme can meet the requirements of the realtime characteristics and the control performance.

Key words: BP neural networks;PID controller;DSP;motor control