北京航空航天大学学报 ›› 2011, Vol. 37 ›› Issue (1): 58-62.

• 论文 • 上一篇    下一篇

资源受限嵌入式机电控制系统功耗感知调度

李锋, 王田苗, 魏洪兴   

  1. 北京航空航天大学 机械工程及自动化学院, 北京 100191
  • 收稿日期:2009-11-11 出版日期:2011-01-31 发布日期:2011-01-28
  • 作者简介:李 锋(1976-),男,山东邹城人,博士生,lifeng@me.buaa.edu.cn,现就职于国家农业信息化工程技术研究中心.
  • 基金资助:

    国家863基金资助项目(2007AA041701)

Power aware scheduling of resource-constraint embedded mechatronic control systems

Li Feng, Wang Tianmiao, Wei Hongxing   

  1. School of Mechanical Engineering and Automation, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2009-11-11 Online:2011-01-31 Published:2011-01-28

摘要: 针对嵌入式机电控制系统的能耗受限问题,提出了一种基于功耗感知和反馈调度的能耗管理方法.研究了嵌入式机电控制系统的静态功耗调度模型,提出了一种基于控制代价和系统能耗的联合优化指标,并给出了其优化解.针对在动态调度模型下运算开销大,难以在线实现的问题,采用智能计算方法逼近优化结果.对比了4种智能计算方法的逼近精度和调度开销,采用小波神经网络逼近优化结果.仿真的结果表明:该方法在保证了控制性能的同时,降低了调度开销和系统能耗.

Abstract: To solve energy constraint of embedded mechatronic control systems(EMCS), an energy management method based on feedback scheduling and power aware was proposed. The static optimization model of EMCS was discussed. A joint optimization target based on the control cost and the system energy consumption was proposed, and the optimal solution was given. In dynamic scheduling, the overhead of static optimization is very high and difficult to realize, so intelligent computing methods were used to approximate the optimal solution. Scheduling overhead and approximation accuracy of four intelligent computing methods were compared, and wavelet neural network was used to approach optimal results. Results of simulations show that the energy consumption on the base of maintaining control performance was reduced effectively by the proposed method.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发