Quantisation-based robust control of uncertain non-strict-feedback non-linear systems under arbitrary switching
Quantisation-based robust control of uncertain non-strict-feedback non-linear systems under arbitrary switching
复制标题
任意切换下不确定非严格反馈非线性系统的基于量化的鲁棒控制
DOI：
10.1049/iet-cta.2015.0679
复制
发表时间：
2016-03
影响因子：
2.6
通讯作者：
Chen Chun Lung Philip
中科院分区：
文献类型：
--
作者：
Lai Guanyu;Liu Zhi;Zhang Yun;Chen Chun Lung Philip
研究了一类非严格反馈非线性切换系统在任意切换下的鲁棒量化控制问题。与非切换非线性系统不同，切换非线性系统（特别是在任意切换下）的稳定性在很大程度上依赖于公共李雅普诺夫函数（CLF）的构造。然而，由于同时考虑全状态不确定性、未知虚拟控制系数以及量化问题，因此在本研究中设计此类CLF相当困难。为了克服这一挑战，本研究提出了一种系统的控制设计方法。具体地说，将具有最小学习参数的神经网络与高斯基函数的有界性相结合来处理系统的全状态不确定性，同时充分利用虚拟控制器的特殊结构来构造在线估计器来处理虚拟系数未知的问题.此外，还提出了一种量化器的非线性分解方法，使得量化问题可以在反推迭代的最后一步独立求解。利用这些结果，成功地建立了一种自适应控制算法，该算法保证了所有子系统的CLF，使得跟踪误差渐近地进入原点的可调区域，同时所有闭环信号在任意切换下保持一致最终有界。最后，仿真结果验证了所得结论的正确性.
This study concentrates on the problem of robust quantised control for a class of switched non-strict-feedback non-linear systems under arbitrary switching. Differs from non-switched non-linear systems, the stability in switched ones (especially under arbitrary switching) is much dependent on the construction of common Lyapunov function (CLF). However, such CLF is fairly difficult to be designed in this study as the simultaneous consideration of full-states uncertainties, unknown virtual control coefficients, as well as quantisation problem. To overcome the challenge, this study presents a systematic control design method. Specifically, the neural networks with minimal learning parameter, and the boundedness property of Gaussian basis functions are combined to deal with the full-states uncertainties, while an online estimator is built by sufficiently using the special structure of virtual controllers to dispose the unknown virtual coefficients problem. Moreover, a non-linear decomposition of quantiser is further proposed, rendering that the quantisation problem can be solved independently at the last step of backstepping iteration. With these results, an adaptive control algorithm which guarantees a CLF for all subsystems is successfully established such that the tracking error is steered into an adjustable area of origin asymptotically, and meanwhile all closed-loop signals remain uniformly ultimately bounded under arbitrary switching. Lastly, the obtained conclusions are well verified by the simulated results.
登录
查看更多内容
DOI：
10.1109/tnnls.2014.2341242
发表时间：
2015-07
影响因子：
10.4
作者：
Lijun Long;Jun Zhao
通讯作者：
Lijun Long;Jun Zhao
DOI：
10.1109/tsmc.2015.2426131
发表时间：
2016
期刊：
IEEE Transactions on Systems, Man, and Cybernetics: Systems
影响因子：
--
作者：
Shaocheng Tong;Lili Zhang;Yong-ming Li
通讯作者：
Shaocheng Tong;Lili Zhang;Yong-ming Li
影响因子：
6.4
作者：
Zhao, Xudong;Zhang, Lixian;Liu, Ming
通讯作者：
Liu, Ming
影响因子：
11.8
作者：
Bing Chen;Chong Lin;Xiaoping Liu;Kefu Liu
通讯作者：
Bing Chen;Chong Lin;Xiaoping Liu;Kefu Liu
DOI：
10.1016/s0005-1098(03)00151-1
发表时间：
2003-09
期刊：
Autom.
影响因子：
--
作者：
D. Liberzon
通讯作者：
D. Liberzon