正在加载图片...
第7期 张蕾等:混合时变时滞神经网络的状态估计器设计 ·851· x+11+1x-1I],非线性干扰取为f(t,x(t))= 1.0 一x的真实值 0.1cosr(t)+0.3.因此,可以得到 0.8片 一-文就101的结果 028小0 0 ·本文结果 1 0.6 -0.05 0.4 01 0.2 显然,lsit|在t=kπ(k=0,±1,±2,…)时不 可微,因此文献1-12]中的判据不能求解此状态 0.2 估计问题.利用Matlab的LMl工具箱,可以求解定 0.46 12345678910 理1得到 P=20.8384 -0.32621 图1状态真实值()和估计值()的响应曲线 1-0.3262 15.4626J Fig.1 Response curves of the true state x (t)and its estimation () 10.4011 -0.17991 Q1= 1-0.1799 7.4702' 0=12.2961 1.0 -0.21441 0.8 一,的真实值 -0.2144 8.7461 0.6 --文献10的结果 ·本文结果 r35.76171.81911 0.4 Q3= 1.8191 33.6083 14.8976 -0.16361 T= -0.2 -0.1636 12.4536' -0.4 6.2605 -0.10631 -0.6 S= 【-0.1063 4.3965J -0.81234567890 -36.7922 R= 0.92661 0.5031 -38.6052 图2状态真实值x2()和估计值()的响应曲线 U= r33.7099 Fig.2 Response curves of the true state x2(t)and its estimation 01 32.2146 2() V=22.9655 01 等式形式给出.数值算例证明了本文设计方法的有 0 19.9233 效性和优越性. W=68.4948 0 0 83.6090' 参考文献 且状态估计增益矩阵为 [1]Liu Y R,Wang Z D,Liu X H.Global exponential stability of gen- -1.7657 0.00541 eralized recurrent neural networks with discrete and distributed de- K= lays.Neural Nettcorks,2006,19 (5):667 -0.0047 -2.4966」 [2]Huang H,Feng G,Cao J D.An LMI approach to delay-dependent 而利用文献0]可以求得 state estimation for delayed neural networks.Neurocomputing, -2.4904 0.00141 2008,71(13H5):2857 K= -0.0114-3.2391 B]He Y,Liu C P,Rees D.New delay-dependent stability criteria 仿真结果可见图1和图2. for neural networks with timevarying delay.IEEE Trans Neural Networks,2007,18(1):310 4结论 4]He Y,Liu G P,Rees D,et al.Stability analysis for neural net- works with timevarying interval delay.IEEE Trans Neural Net- 本文主要研究具有混合时变时滞神经网络的状 Work,2007,18(6):1850 态估计问题.离散时滞是时变的,并且在一个下界 [5]Gao H J,Chen T W.New results on stability of discretetime sys- 不一定为零的区间内变化.通过定义新的Lyapunov tems with time-varying state delay.IEEE Trans Autom Control, 2007,52(2):328 泛函,结合Jensen积分不等式,确定了确保误差状 [6]Shao H Y.New delay-dependent stability criteria for systems with 态系统全局渐近稳定的充分条件,并且得到系统状 interval delay.Automatica,2009,45 (3):744 态估计器的设计方法.所得结果完全由线性矩阵不 7]Wang Z D,Ho D W C,Liu X H.State estimation for delayed第 7 期 张 蕾等: 混合时变时滞神经网络的状态估计器设计 [| x + 1 | + | x - 1 | ],非线性干扰取为 f( t,x( t) ) = 0. 1cosx( t) + 0. 3. 因此,可以得到 Σ1 = - 0. 2 0 [ ] 0 - 0. 2 ,Σ2 = - 0. 05 0 [ ] 0 - 0. 05 , Σ3 = 0. 08 0 [ ] 0 0. 08 ,Σ4 = 0. 3 0 [ ] 0 0. 3 . 显然,|sint | 在 t = kπ( k = 0,± 1,± 2,…) 时不 可微,因此文献[11--12]中的判据不能求解此状态 估计问题. 利用 Matlab 的 LMI 工具箱,可以求解定 理 1 得到 P = 20. 838 4 - 0. 326 2 [ ] - 0. 326 2 15. 462 6 , Q1 = 10. 401 1 - 0. 179 9 [ ] - 0. 179 9 7. 470 2 , Q2 = 12. 296 1 - 0. 214 4 [ ] - 0. 214 4 8. 746 1 , Q3 = 35. 761 7 1. 819 1 [ ] 1. 819 1 33. 608 3 , T = 14. 897 6 - 0. 163 6 [ ] - 0. 163 6 12. 453 6 , S = 6. 260 5 - 0. 106 3 [ ] - 0. 106 3 4. 396 5 , R = - 36. 792 2 0. 926 6 [ ] 0. 503 1 - 38. 605 2 , U = 33. 709 9 0 [ ] 0 32. 214 6 , V = 22. 965 5 0 [ ] 0 19. 923 3 , W = 68. 494 8 0 [ ] 0 83. 609 0 , 且状态估计增益矩阵为 K = - 1. 765 7 0. 005 4 [ ] - 0. 004 7 - 2. 496 6 . 而利用文献[10]可以求得 K = - 2. 490 4 0. 001 4 [ ] - 0. 011 4 - 3. 239 1 . 仿真结果可见图 1 和图 2. 4 结论 本文主要研究具有混合时变时滞神经网络的状 态估计问题. 离散时滞是时变的,并且在一个下界 不一定为零的区间内变化. 通过定义新的 Lyapunov 泛函,结合 Jensen 积分不等式,确定了确保误差状 态系统全局渐近稳定的充分条件,并且得到系统状 态估计器的设计方法. 所得结果完全由线性矩阵不 图 1 状态真实值 x1 ( t) 和估计值 x^ 1 ( t) 的响应曲线 Fig. 1 Response curves of the true state x1 ( t) and its estimation x^ 1 ( t) 图 2 状态真实值 x2 ( t) 和估计值 x^ 2 ( t) 的响应曲线 Fig. 2 Response curves of the true state x2 ( t) and its estimation x^ 2 ( t) 等式形式给出. 数值算例证明了本文设计方法的有 效性和优越性. 参 考 文 献 [1] Liu Y R,Wang Z D,Liu X H. Global exponential stability of gen￾eralized recurrent neural networks with discrete and distributed de￾lays. Neural Networks,2006,19( 5) : 667 [2] Huang H,Feng G,Cao J D. An LMI approach to delay-dependent state estimation for delayed neural networks. Neurocomputing, 2008,71( 13-15) : 2857 [3] He Y,Liu G P,Rees D. New delay-dependent stability criteria for neural networks with time-varying delay. IEEE Trans Neural Networks,2007,18( 1) : 310 [4] He Y,Liu G P,Rees D,et al. Stability analysis for neural net￾works with time-varying interval delay. IEEE Trans Neural Net￾works,2007,18( 6) : 1850 [5] Gao H J,Chen T W. New results on stability of discrete-time sys￾tems with time-varying state delay. IEEE Trans Autom Control, 2007,52( 2) : 328 [6] Shao H Y. New delay-dependent stability criteria for systems with interval delay. Automatica,2009,45( 3) : 744 [7] Wang Z D,Ho D W C,Liu X H. State estimation for delayed ·851·
<<向上翻页向下翻页>>
©2008-现在 cucdc.com 高等教育资讯网 版权所有