正在加载图片...
赵栎等:确定性多变量自校正控制的稳定性、收敛性和鲁棒性 ·1221· 系统动态性能的影响相当于增加了一个扰动信号 (5):879 e(k)(或者折算为△u(k)),这是自适应或自学习所 [6]Goodwin G C.Sin K S.Adaptive Filtering Prediction and Control 付出的必然代价.对系统的稳定性、收敛性指标来 Courier Corporation,2014 [7]Lozano-Leal R,Goodwin G.A globally convergent adaptive pole 说,在一定条件下这一扰动信号没有造成影响,但对 placement algorithm without a persistency of excitation require- 系统的动态指标的影响是明显的 ment.IEEE Trans Autom Control,1985,30(8):795 有3种方法可以用来抑制e(k)/△u(k)的影 [8]Lozano-Leal R.Robust adaptive regulation without persistent exci- 响:第一,采用多模型自适应控制的方法,在e(k)较 tation.IEEE Trans Autom Control,1989,34(12)1260 大时,采用基于先验知识的固定模型:第二,修改控 [9]Lozano R.Singularity-free adaptive pole-placement without resor- ting to persistency of excitation:detailed analysis for first order 制策略,即对参数估计误差进行预测,并考虑到控制 systems.Automatica,1992,28(1):27 律中,这种思路和自适应预测控制殊途同归:第三, [10]Lozano R,Dion J M,Dugard L Singularity-free adaptive pole 因为产生△u(k)的根源是参数估计误差e(k),而从 placement using periodic controllers.IEEE Trans Autom Control, e(k)到△u(k)的映射由控制策略决定,因此可以通 1993,38(1):104 过采用鲁棒控制策略来尽量抑制e(k). [11]Lozano R.Zhao X H.Adaptive pole placement without excitation probing signals.IEEE Trans Autom Control,1994,39(1):47 4结论 [12]Xu LZ,Wang X H.Methods and Examples of Mathematical A- nalysis (Revised Edition).Beijing:Higher Education Press, 借助虚拟等价系统,将自校正控制系统的稳定 1983 性、收敛性和鲁棒性分析的困难从系统结构上转移 (徐利治,王兴华.数学分析的方法及例题选讲(修订版) 到补偿信号上,从而降低了原问题的难度,使得自校 北京:高等教育出版社,1983) 正控制系统的稳定性、收敛性乃至鲁棒性分析变得 [13]Caines P,Lafortune S.Adaptive control with recursive identifica- tion for stochastic linear systems.IEEE Trans Autom Control, 更加直观、易于理解。得到的若干定理和推论,在一 1984,29(4):312 定程度上可以统一对自校正控制系统的稳定性和收 [14]Chen H F,Guo L.Asymptotically optimal adaptive control with 敛性的判断和认识,同时得出一些新的观点,包括: consistent parameter estimates.SIAM Control Optim,1987,25 参数估计的收敛性不是自校正控制系统稳定和收敛 (3):558 的必要条件:系统自身的反馈信息对自校正控制系 [15]Astrom K J,Wittenmark B.Adaptire Control.Newyork:Courier 统的控制目的是充分的,即,外加激励信号不是必 Corporation,2013 [16]Hu S G.Functional Analysis.Beijing:Higher Education Press, 须的 2007 (胡适耕.泛函分析.北京:高等教育出版社,2007) 参考文献 [17]Prandini M,Campi M C.A new recursive identification algorithm [1]Zhang WC.Research on Robust Adaptive Control Theory and Its for singularity free adaptive control.Syst Control Lett,1998,34 Applications[Dissertation].Beijing:Tsinghua University,1993 (4):177 (张维存。鲁棒自适应控制理论及应用研究[学位论文].北 [18]Prandini M,Bittanti S,Campi M C.A penalized identification 京:清华大学,1993) criterion for securing controllability in adaptive control.Math [2]Fekri S,Athans M,Pascoal A.Issues,progress and new results Syst Estim Control,1998,8(4):1 in robust adaptive control.Int J Adapt Control Signal Process, [19]Liberzon D,Morse A S.Basic problems in stability and design of 2006,20(10):519 switched systems.IEEE Control Syst Mag,1999,19(5):59 [3]Li QQ.Adaptire Control System:Theory,Design,and Applica- [20]Shorten R,Wirth F,Mason 0,et al.Stability criteria for tions.Beijing:Science Press,1990 switched and hybrid systems.SIAM Rev,2007,49(4):545 (李清泉.自适应控制系统理论、设计与应用.北京:科学出 [21]Feng C B,Shi W.Adaptire control.Beijing:Publishing House 版社,1990) of Electronics Industry,1986 [4]Xie X M,Ding F.Adaptice Control System.Beijing:Tsinghua U- (冯纯伯,史维.自适应控制,北京:电子工业出版社,1986) niversity Press,2002 [22]Zhang W C,Wei W.Virtual equivalent system theory for adap- (谢新民.丁锋.自适应控制系统.北京:清华大学出版社, tive control and simulation verification.Sci Sin Inform,2018,48 2002) (7):947 [5]Zhang W C.On the stability and convergence of self-tuning con- (张维存,魏伟.自适应控制的虚拟等价系统理论及仿真验 trol-virtual equivalent system approach.Int Control,2010,83 证.中国科学:信息科学,2018,48(7):947)赵 栎等: 确定性多变量自校正控制的稳定性、收敛性和鲁棒性 系统动态性能的影响相当于增加了一个扰动信号 e(k)(或者折算为 驻u(k)),这是自适应或自学习所 付出的必然代价. 对系统的稳定性、收敛性指标来 说,在一定条件下这一扰动信号没有造成影响,但对 系统的动态指标的影响是明显的. 有 3 种方法可以用来抑制 e( k) / 驻u( k) 的影 响:第一,采用多模型自适应控制的方法,在 e(k)较 大时,采用基于先验知识的固定模型;第二,修改控 制策略,即对参数估计误差进行预测,并考虑到控制 律中,这种思路和自适应预测控制殊途同归;第三, 因为产生 驻u(k)的根源是参数估计误差 e(k),而从 e(k)到 驻u(k)的映射由控制策略决定,因此可以通 过采用鲁棒控制策略来尽量抑制 e(k). 4 结论 借助虚拟等价系统,将自校正控制系统的稳定 性、收敛性和鲁棒性分析的困难从系统结构上转移 到补偿信号上,从而降低了原问题的难度,使得自校 正控制系统的稳定性、收敛性乃至鲁棒性分析变得 更加直观、易于理解. 得到的若干定理和推论,在一 定程度上可以统一对自校正控制系统的稳定性和收 敛性的判断和认识,同时得出一些新的观点,包括: 参数估计的收敛性不是自校正控制系统稳定和收敛 的必要条件;系统自身的反馈信息对自校正控制系 统的控制目的是充分的,即,外加激励信号不是必 须的. 参 考 文 献 [1] Zhang W C. Research on Robust Adaptive Control Theory and Its Applications [Dissertation]. Beijing: Tsinghua University, 1993 (张维存. 鲁棒自适应控制理论及应用研究[学位论文]. 北 京: 清华大学, 1993) [2] Fekri S, Athans M, Pascoal A. Issues, progress and new results in robust adaptive control. Int J Adapt Control Signal Process, 2006, 20(10): 519 [3] Li Q Q. Adaptive Control System: Theory, Design, and Applica鄄 tions. Beijing: Science Press, 1990 (李清泉. 自适应控制系统理论、设计与应用. 北京: 科学出 版社, 1990) [4] Xie X M, Ding F. Adaptive Control System. Beijing: Tsinghua U鄄 niversity Press, 2002 (谢新民, 丁锋. 自适应控制系统. 北京: 清华大学出版社, 2002) [5] Zhang W C. On the stability and convergence of self鄄tuning con鄄 trol鄄virtual equivalent system approach. Int J Control, 2010, 83 (5): 879 [6] Goodwin G C, Sin K S. Adaptive Filtering Prediction and Control. Courier Corporation, 2014 [7] Lozano鄄Leal R, Goodwin G. A globally convergent adaptive pole placement algorithm without a persistency of excitation require鄄 ment. IEEE Trans Autom Control, 1985, 30(8): 795 [8] Lozano鄄Leal R. Robust adaptive regulation without persistent exci鄄 tation. IEEE Trans Autom Control, 1989, 34(12): 1260 [9] Lozano R. Singularity鄄free adaptive pole鄄placement without resor鄄 ting to persistency of excitation: detailed analysis for first order systems. Automatica, 1992, 28(1): 27 [10] Lozano R, Dion J M, Dugard L. Singularity鄄free adaptive pole placement using periodic controllers. IEEE Trans Autom Control, 1993, 38(1): 104 [11] Lozano R, Zhao X H. Adaptive pole placement without excitation probing signals. IEEE Trans Autom Control, 1994, 39(1): 47 [12] Xu L Z, Wang X H. Methods and Examples of Mathematical A鄄 nalysis ( Revised Edition ). Beijing: Higher Education Press, 1983 (徐利治, 王兴华. 数学分析的方法及例题选讲(修订版). 北京: 高等教育出版社,1983) [13] Caines P, Lafortune S. Adaptive control with recursive identifica鄄 tion for stochastic linear systems. IEEE Trans Autom Control, 1984, 29(4): 312 [14] Chen H F, Guo L. Asymptotically optimal adaptive control with consistent parameter estimates. SIAM J Control Optim, 1987, 25 (3): 558 [15] 魡str觟m K J, Wittenmark B. Adaptive Control. Newyork: Courier Corporation, 2013 [16] Hu S G. Functional Analysis. Beijing: Higher Education Press, 2007 (胡适耕. 泛函分析. 北京: 高等教育出版社, 2007) [17] Prandini M, Campi M C. A new recursive identification algorithm for singularity free adaptive control. Syst Control Lett, 1998, 34 (4): 177 [18] Prandini M, Bittanti S, Campi M C. A penalized identification criterion for securing controllability in adaptive control. J Math Syst Estim Control, 1998, 8(4): 1 [19] Liberzon D, Morse A S. Basic problems in stability and design of switched systems. IEEE Control Syst Mag, 1999, 19(5): 59 [20] Shorten R, Wirth F, Mason O, et al. Stability criteria for switched and hybrid systems. SIAM Rev, 2007, 49(4): 545 [21] Feng C B, Shi W. Adaptive control. Beijing: Publishing House of Electronics Industry,1986 (冯纯伯,史维. 自适应控制,北京:电子工业出版社,1986) [22] Zhang W C, Wei W. Virtual equivalent system theory for adap鄄 tive control and simulation verification. Sci Sin Inform, 2018, 48 (7): 947 (张维存, 魏伟. 自适应控制的虚拟等价系统理论及仿真验 证. 中国科学: 信息科学, 2018, 48(7): 947) ·1221·
<<向上翻页
©2008-现在 cucdc.com 高等教育资讯网 版权所有