Energy dissipating hybrid control for impulsive dynamical systems (Q955632): Difference between revisions
From MaRDI portal
Latest revision as of 19:46, 28 June 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Energy dissipating hybrid control for impulsive dynamical systems |
scientific article |
Statements
Energy dissipating hybrid control for impulsive dynamical systems (English)
0 references
20 November 2008
0 references
The authors consider controlled impulsive dynamical systems of the form \[ \dot{x}_p(t)=f_{cp}(x_p(t),u_c(t)),\quad x_p(0)=x_{p0},\quad (x_p(t),u_c(t))\not \in \mathcal{Z}_p, \] \[ \Delta x_p(t)=f_{dp}(x_p(t),u_d(t)),\quad (x_p(t),u_c(t)) \in \mathcal{Z}_p, \] \[ y(t)=h_p(x_p(t)), \] where \(t \geq 0\), \(x_p(t) \in \mathcal{D}_p \subseteq \mathbb{R}^{n_p}\), \( \mathcal{D}_p \) is an open set containing \(0\), \(\Delta x_p(t)= x_p(t^+)-x_p(t)\), \(u_c(t) \in \mathbb{R}^{m_c}\), \(u_d(t) \in \mathbb{R}^{m_d}\), \(f_{c_p}: \mathcal{D}_p \times \mathbb{R}^{m_c} \longrightarrow \mathbb{R}^{n_p}\) is infinitely differentiable and \(f_{dp}:\mathcal{D}_p\times \mathbb{R}^{m_d} \longrightarrow \mathbb{R}^{n_p}\) is continuous, \(h_p:\mathcal{D}_p \longrightarrow \mathbb{R}^l\) is continuous (additional hypotheses have to be imposed on these functions) and \(\mathcal{Z}_p=\mathcal{Z}_{x_p} \times \mathcal{Z}_{u_c} \subset \mathcal{D}_p \times \mathbb{R}^{m_c}\) is the resetting set. Besides, they also consider \(n_c\)-dimensional hybrid (resetting) dynamic controllers of the type \[ \dot{x}_c(t)=f_{cc}(x_c(t),y(t)),\quad x_c(0)=x_{c0},\quad (x_c(t),y(t))\not \in \mathcal{Z}_c, \] \[ \Delta x_c(t)=f_{dc}(x_c(t),y(t)),\quad (x_c(t),y(t)) \in \mathcal{Z}_c, \] \[ u_c(t)=h_{cc}(x_c(t),y(t)), \] \[ u_d(t)=h_{dc}(x_c(t),y(t)), \] for \(t \geq 0\), \(x_c(t) \in \mathcal{D}_c \subseteq \mathbb{R}^{n_c}\), \(\mathcal{D}_c\) an open set containing \(0\), \(f_{cc}:\mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{n_c}\) smooth and \(f_{dc}: \mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{n_c}\), \(h_{cc}:\mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{m_c}\), \(h_{dc}:\mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{m_d}\) continuous satisfying the appropriate hypotheses, and \(\mathcal{Z}_c \subset \mathcal{D}_c \times \mathbb{R}^l\) the resetting set. These dynamical systems are written as hybrid closed-loop systems given by the continuous-time dynamics \[ \dot{x}(t)=f_{c}(x(t)),\quad x(0)=x_{0},\quad x(t)\not \in \mathcal{Z}, \] combined with the resetting law \[ \Delta x(t)=f_{d}(x(t)),\quad x(t)\in \mathcal{Z}. \] Sufficient conditions are provided in order to guarantee the asymptotical stability of the zero solution by calculating a subset of its domain of attraction. It is also studied the following hybrid controller for lossless impulsive dynamical systems \[ \dot{x}_c(t)=f_{cc}(x_c(t),y(t)),\quad x_c(0)=x_{c0},\quad (x_c(t),y(t))\not \in \mathcal{Z}_c, \] \[ \Delta x_c(t)=\eta(y(t))-x_c(t),\quad (x_c(t),y(t)) \in \mathcal{Z}_c, \] \[ y_{cc}(t)=h_{cc}(x_c(t),u_{cc}(t)), \] \[ y_{dc}(t)=h_{dc}(x_c(t),y(t)), \] where \(x_c(t) \in \mathcal{D}_c \subseteq \mathbb{R}^{n_c}\), \(\mathcal{D}_c\) is an open set containing \(0\), \(y(t) \in \mathbb{R}^l\), \(y_{cc}(t) \in \mathbb{R}^{m_c}\), \(y_{dc}(t) \in \mathbb{R}^{m_d}\), \(f_{cc}:\mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{n_c}\) is smooth, and \(\eta :\mathbb{R}^l \longrightarrow \mathcal{D}_c\), \(h_{cc}: \mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{m_c}\), \(h_{dc}:\mathcal{D}_c \times \mathbb{R}^l \longrightarrow \mathbb{R}^{m_d}\) are continuous satisfying the appropriate properties. Sufficient conditions are also given for asymptotic stability of the closed-loop system using hybrid controllers. Impulsive port-controlled Hamiltonian systems are studied, presenting a fixed-order energy-based hybrid controller. Besides, a hybrid control framework is presented for nonsmooth Euler-Lagrange dynamical systems and sufficient conditions are provided for its stabilization. Finally, a constrained inverted pendulum is analyzed by considering an hybrid dynamic compensator to stabilize the equilibrium point. The proposed approach differs from energy-based controllers for continuous-time systems in the method used to achieve stabilization.
0 references
Hybrid control
0 references
Hybrid systems
0 references
Dynamic compensation
0 references
Impulsive dynamical systems
0 references
Lossless systems
0 references
Nonsmooth Euler-Lagrange systems
0 references
0 references
0 references
0 references
0 references