Multi-Pulse-Width Modulated Control of Linear Systems

From MaRDI portal
Publication:3843022





This paper deals with the multi-pulse-width modulated (MPWM) control of a time-invariant linear system modeled by \[ \dot{\mathbf x}= A{\mathbf x}+B{\mathbf u},tag 1 \] where \({\mathbf x}= [x_j(t)]_{n\times 1}\), \({\mathbf u}= [u_j(t)]_{m\times 1}\), \[ u_j(t)= u_j^{(\text{PWM})}(t)= \begin{cases} u_{Mj}, &\text{if }k\Delta+ T_{ji}- \delta_{ji}\leq t\leq k\Delta+ T_{ji},\\ 0, &\text{otherwise} \end{cases} \tag{2} \] and \(\Delta\) is the sampling interval. The study is based on the comparison between the MPWM control and its parent discrete time control -- called pulse amplitude modulation (PAM): \[ u_j(t)= u_j^{(\text{PAM})}(t)= u_j(k\Delta), \quad k\Delta\leq t\leq (k+1)\Delta. \] For a single-input system with \(B={\mathbf b}\), the residual error between the two responses is \[ \varepsilon_k= \sum_{j=2p}^\infty \frac{A^{j+1}{\mathbf b}} {(j+1)!} \Biggl\{ \Delta^{j+1} u_k- \sum_{i=1}^p [(\Delta- T_i+\delta_i)^{j+1}- (\Delta- T_i)^{j+1}] u_M \Biggr\}, \tag{4} \] which is a decreasing function of the number of pulses \(p\) per sampling interval \(\Delta\). Thus, a method is presented to convert a parent discrete time control into an MPWM equivalent, and such an equivalence can be established by matching the system state response at each sampling interval. Moreover, provided a sufficiently large \(p\) is used, the degree of approximation can be as high as desired even if a relatively slow sampling frequency is adopted. A numerical example on a simple fourth-order system provides the basis for understanding the pros and cons of the method, which is then applied to a more complex aerospace problem: the stabilization of coupled thermoelastic vibrations of a space structure.











This page was built for publication: Multi-Pulse-Width Modulated Control of Linear Systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3843022)