Robust stabilization and ultimate boundedness of dynamic surface control systems via convex optimization (Q4783548)

The authors develop a procedure for choosing the controller gains and analysing the stability of a class of nonlinear systems controlled via dynamic surface control (DSC) in the presence of modelling uncertainty in the strict feedback form \[ \left\{\begin{aligned} \dot x_1&= x_2 +f_1(x_1) + \Delta f_1(x_1)\\ \dot x_2&= x_3 +f_2(x_1,x_2) + \Delta f_2(x_1,x_2)\\ &\dots\\ \dot x_{n-1}&=x_n + f_{n-1}(x_1,\dots,x_{n-1}) + \Delta f_{n-1}(x_1,\dots, x_{n-1})\\ \dot x_n&=u.\end{aligned}\right. \] The `strict-feedback' refers to the fact that the nonlinearities \(f_i\) and \(\Delta f_i\) depend only on \(x_1,\dots,x_i\). In addition, the nonlinearities are assumed to satisfy the following properties. First, \(f_i\) is a sufficiently smooth function in its arguments, and \(f_i(0,\dots,0) = 0\) for \(i =1,\dots,n-1\). Second, \(\Delta f_i\) is an unknown non-Lipschitz nonlinearity, but it is bounded by a known \(C^1\) function \(\rho_i\) in a domain \(\mathcal D =\{x\in\mathbb R^n: \|x\|\leq d_0, d_0>0\}\), such that \(|\Delta_if(x_1,\dots,x_i)|\leq\rho_i(x_1,\dots,x_i)\) and \(\rho_i(0,\dots,0) = 0\). Furthermore, \(\Delta_i f\) is also a \(C^0\) function to guarantee the existence of solutions. The DSC provided linear error dynamics with perturbation terms as a function of the error. This fact is used to assign the desired eigenvalues to the closed-loop error system matrix, while satisfying the DSC constraints on the surface gains and time constants. When a bound for the uncertainties is assumed or a bound on the reference input is defined for a tracking problem, the authors perform the stability test with respect to the controller gains and filter time constants determined in the previous step, and determine the robustness of the fixed controller by solving a linear matrix inequality. Finally, an ellipsoidal approximation of the tracking error bounds for a tracking problem is obtained via convex optimization.