scientific article; zbMATH DE number 1293746

From MaRDI portal
Publication:4246029





This paper deals with the output regulation of general uncertain single-variable nonlinear systems in the form of \[ \begin{aligned} \dot x_i & = x_{i+1}+ \theta^\tau \varphi_i(x_1,\dots, x_i)+ \psi_i({\mathbf x},t),\quad 1\leq i\leq n-1,\\ \dot x_n & = \varphi_0({\mathbf x})+ \theta^\tau \varphi_n({\mathbf x})+ \psi_n({\mathbf x},t)+ \beta_0({\mathbf x}) u,\\ y & =x_1,\end{aligned} \] where \({\mathbf x}\in\mathbb{R}^n\), \(u\), \(y\) are respectively the state vector, control and output; \(\theta\in R\) unknown constant vector; \(\varphi_i(\cdot)\), \(0\leq i\leq n\) known smooth nonlinear functions; \(\beta_0({\mathbf x})\neq 0\), \(\forall{\mathbf x}\in\mathbb{R}^n\); \(\psi_i({\mathbf x},t)\leq \lambda_i|\mu_i({\mathbf x},t)|\), \(\lambda_i>0\), \(1\leq i\leq n\) reflect the system uncertainties (including nonlinear uncertainties, bounded disturbances and time-varying parameters), and \(\lambda_i\) may be unknown, \(\mu_i({\mathbf x},t)\) is smooth nonlinear. Using a recurrence design approach, i.e. letting \[ \dot\theta_{i-1}= z_{i- 1}\Biggl(\varphi_{i- 1}- \sum^{i- 2}_{j= 1} {\partial\alpha_{i-2}\over \partial x_j} \varphi_j\Biggr)- l\theta_{i-1}\quad\text{and} \quad z_i= x_i- \alpha_{i-1}(x_{i- 1}) \] such that \(\dot z_i= z_{i+1}+ \alpha_i+ \theta^\tau \varphi_i+ \psi_i\), \(1\leq i\leq n-1\), a new class of robust adaptive controllers for output regulation is obtained \[ u=-{1\over \beta_0} \Biggl[c_n z_n+ z_{n-1}+ \varphi_0+ \theta^\tau_n \Biggl(\varphi_n- \sum^{n-1}_{j=1} {\partial\alpha_{n-1}\over\partial x_j} \varphi_j\Biggr)- \] \[ -\sum^{n-1}_{j=1} \Biggl({\partial\alpha_{n- 1}\over\partial x_j} x_{j+ 1}+ {\partial\alpha_{n- 1}\over\partial\theta_j} \dot\theta_j\Biggr)+K_n z_nu^2_n+ K_n z_n \sum^{n-1}_{j= 1}\Biggl({\partial\alpha_{n- 1}\over\partial x_j} u_j\Biggr)^2\Biggr], \] \[ \dot\theta_n= z_n\Biggl(\varphi_n- \sum^{n- 1}_{j= 1}{\partial\alpha_{n-1}\over\partial x_j} \varphi_j\Biggr)- {\l}\theta_n. \] Via the Lyapunov function method, the controllers can be proved to ensure the global boundedness of all closed-loop signals. It has the advantage that less knowledge of the uncertainties is required and, therefore, can deal with much more system uncertainties than the existing schemes in this area. The result obtained here much further extends the result about parametric-strict-feedback systems with bounded disturbances. A simulation example shows the good control effects of the designed adaptive controller.




Cited in
(36)








This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4246029)