For the linear quadratic Gaussian (LQG) problems with unknown constant parameters and accurately measurable state, the separation theorem is no longer valid due to the coupling between parameters estimation and control gain, which lead to the failure to the analytical solution of the control law. The dual adaptive control method is proposed in this paper, where a state space model of parameters estimation is established, control gain is obtained by rolling dynamic programming, Kalman flter is used to estimate unknown parameters, the present difculties about the mutual coupling between estimated gain and control gain are overcome, and the controller with suboptimal characteristics is designed. On the one hand, the controller can optimize the control target, on the other hand, it can also learn unknown parameters effectively. The simulation results show the effectiveness of the proposed control algorithm.
[1] Xie L L, Guo L. How much uncertainty can be dealt with by feedback[J]. IEEE Transactions on Automatic Control, 2000, 45(12):2203-2220.
[2] 韩崇昭. 随机系统概论——分析、估计与控制[M]. 北京:清华大学出版社,2014.
[3] 俞立. 鲁棒控制——线性矩阵不等式处理方法[M]. 北京:清华大学出版社,2002.
[4] Yuri B S, Jaime A M, Leonid M F. Twisting sliding mode control with adaptation:Lyapunov design, methodology and application[J]. Automatica, 2017, 75(1):229-235.
[5] Saha B, Goebel K, Poll S. Prognostics methods for battery health monitoring using a Bayesian framework[J]. IEEE Transactions on Instrumentation and Measurement, 2009, 58(2):291-296.
[6] Bania P. Simple example of dual control problem with almost analytical solution[C]//Proceeding of 19th Polish Control Conference, Krakow, Poland, 2017:55-64.
[7] Alpcan T, Shanmes I. An information-based learning approach to dual control[J]. IEEE Transactions on Neural Networks and Learning Systems, 2015, 26(11):2736-2748.
[8] Alpcan T., Shanmes I, Cantoni M, Nair G. Learning and information for dual control[C]//20139th Asian Control Conference (ASCC), 2013:1-6.
[9] Tor A N, Heirunf B, Erik Y, Bharne F. Dual adaptive model predictive control[J]. Automatica, 2017, 80(1):340-348.
[10] Huang J R, Qian F C, Xie G, Yang H Z. Robust learning control for dynamic systems with mixed uncertainties[J]. Journal of Systems Engineering and Electronics,2016, 27(3):656-663.
[11] Li D, Qian F C, Fu P L. Variance minimization approach for a class of dual control problems[J]. IEEE Transactions Automatic Control, 2002, 47(12):2010-2020.
[12] Qian F C, Huang J R, Liu D, Hu S L. Adaptive dual control of discrete-time LQG problems with unknown-but-bounded parameter[J]. Asian Journal of Control, 2015, 17(3):942-951.
[13] 杨恒占,钱富才,黄姣茹,高嵩. 一类随机系统完全统计特征控制[J]. 控制理论与应用,2016, 33(5):669-675. Yang H Z, Qian F C, Huagn J R, Gao S. The complete statistical characterization control for a class of stochastic systems[J]. Control Theory & Applications, 2016, 33(5):669-675. (in Chinese)
[14] Li D, Qian F C, Fu P L. Optimal nominal dual control for discrete-time LQG problem with unknown parameters[J]. Automatica, 2008, 44(1):119-127.
[15] 尚婷,钱富才,张晓艳,谢国. 具有学习特点的控制器设计算法[J]. 自动化学报,2017, 43(7):1202-1207. Shang T, Qian F C, Zhang X Y, Xie G. Controller design algorithm with learning property[J]. Acta Automatica Sinical, 2017, 43(7):1202-1207. (in Chinese)
[16] 尚婷,钱富才,张晓艳,谢国. 具有未知参数的LQG对偶控制算法研究[J]. 自动化学报,2017, 43(8):1474-1480. Shang T, Qian F C, Zhang X Y, Xie G. Research on dual control algorithm for LQG with unknown parameters[J]. Acta Automatica Sinical, 2017, 43(8):1474-1480. (in Chinese)
[17] Qian F C, Zhu S P, Liu D. On LQG problems with unknown noises[J]. Control Theory & Applications, 2010, 27(8):1017-1022.
[18] Huang M, Wang X, Wang Z L. Multiple model adaptive control for a class of linear-bounded nonlinear systems[J]. IEEE Transactions on Automatic Control, 2015, 60(1):271-276.
[19] Ma Y J, Jiang B, Tao G, Cheng Y H. Uncertainty decomposition-based fault-tolerant adaptive control of flexible spacecraft[J]. IEEE Transactions on Aerospace and Electronic Systems, 2015, 51(2):1053-1068.
[20] De V J P, Laskey K, Jousselme A L, Blasch E, Pavlin G, Costa P. Uncertainty representation, quantifcation and evaluation for data and information fusion[C]//Proceedings of the 18th International Conference on Information Fusion. Washington D. C., USA:IEEE, 2015:50-57.
[21] Qian F C, Gao J J, Li D. Complete statistical characterization of discrete-time LQG and cumulant control[J]. IEEE Transactions on Automatic Control, 2012, 57(8):2110-2115.
[22] Feldbaum A A. Dual control theory I-Ⅱ[J]. Automatic Remote Control, 1960, 21(4):1033-1039.
[23] Deshpande J G, Upadhyay T N, Lainiotis D G. Adaptive control of linear stochastic systems[J]. Automatica, 1973, 9(1):107-115.
[24] Wang L Z, Qian F C. Shaping PDF of the state variable based on piecewise linear control for non-linear stochastic systems[J]. Science China:Information Sciences, 2016, 59(11):106-116.
