A Statistical Model for Computational Delays in Digital Control Systems
Although the speed of digital devices has been improved dramatically in recent years, computational delay should still be taken into account in digital control system modeling, analysis and design especially when fast sampling and extensive computation is required. Whatever method is employed to compensate for this delay, its magnitude should be known or estimated. By many experiments, we found that the computational delay has a finite interval and inevitable uncertainty without obviously normal distribution behavior and the four-parameter beta distribution is a better statistical model for the computational delay data in digital control systems. Subsequently, we proposed a fast empirical parameters estimation algorithm. In addition, how to test the models goodness-of-fit was also rendered.
computational delay four-parameter Beta distribution model parameter estimation goodness-of-fit test
Yonlin Zhang Aiguo Song
School of Electronics and Information Jiangsu University of Science and Technology Zhenjiang, Jiangs School of Instrument Science and Engineering Southeast University Nanjing, Jiangsu Province, China
国际会议
2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)
上海
英文
36-40
2010-12-25（万方平台首次上网日期，不代表论文的发表时间）