Open in App
Log In Start studying!

Select your language

Suggested languages for you:

Assume that \(\Sigma\) is observable and that $$ \delta(\lambda-\mu) \neq 2 k \pi i, \quad k=\pm 1, \pm 2, \ldots, $$ for every two eigenvalues \(\lambda, \mu\) of \(A\). Then \(\Sigma\) is also \(\delta\)-sampled observable.

Short Answer

Expert verified
Under the given condition \(\delta(\lambda-\mu) \neq 2 k \pi i\) for every two eigenvalues \(\lambda, \mu\) of \(A\), the continuous-time system \(\Sigma\) is not only observable but also \(\delta\)-sampled observable. This follows from the fact that the observability matrix \(O\) has full rank, and none of the eigenvalues of the system matrix \(A\) will cause the \(e^{A\delta}\) term to become singular or reducible to trivial cases after sampling, ensuring that the observability matrix associated with the discrete-time system \(O_d\) also has full rank.
See the step by step solution

Step by step solution

Unlock all solutions

Get unlimited access to millions of textbook solutions with Vaia Premium

Over 22 million students worldwide already upgrade their learning with Vaia!

Step 1: Define observability and \(\delta\)-sampled observability

Observability is a property of a system that determines whether the system's internal states can be determined by observing its output. In the context of linear systems, a system is observable if we can uniquely determine the initial state of the system from the output. \(\delta\)-sampled observability is a property related to discrete-time systems that concerns whether the system's states can be determined by observing a sampled output, with \(\delta\) being the sampling interval. In other words, instead of having continuous outputs, we only observe the system outputs at specific, discrete time instants.

Step 2: Show that the given condition ensures \(\delta\)-sampled observability

To show that \(\Sigma\) is \(\delta\)-sampled observable under the given condition, we will first recall the result from observability theory that states: A continuous-time LTI (Linear Time-Invariant) system with state-space representation: \[ \begin{cases} \dot{x}(t) = Ax(t) + Bu(t) \\ y(t) = Cx(t) + Du(t) \end{cases}, \] is observable if and only if the observability matrix \(O\) has full rank, where \(O = \begin{bmatrix} C \\ CA \\ \vdots \\ CA^{n-1} \end{bmatrix}\), and \(n\) is the dimension of the state vector \(x(t)\). To show \(\delta\)-sampled observability, we need to consider the discrete-time version of the system, by sampling the output at discrete time intervals \(\delta\): \[ \begin{cases} x[k+1] = e^{A\delta} x[k] + B_d u[k] \\ y[k] = Cx[k] + Du[k] \end{cases}, \] where \(x[k] = x(k\delta)\), \(y[k] = y(k\delta)\), and \(B_d = \int_0^{\delta} e^{As}B ds\). Now, \(\Sigma\) is \(\delta\)-sampled observable if and only if the observability matrix associated with the discrete-time system: \(O_d = \begin{bmatrix} C \\ Ce^{A\delta} \\ \vdots \\ Ce^{A(n-1)\delta} \end{bmatrix}\), has full rank. Given that the continuous-time system is observable, which implies that \(O\) has full rank, and the condition \(\delta(\lambda-\mu) \neq 2 k \pi i\) for every two eigenvalues \(\lambda, \mu\) of \(A\), we can conclude that \(O_d\) will also have full rank. This is because none of the eigenvalues of the system matrix \(A\) will cause the \(e^{A\delta}\) term to become singular or reducible to trivial cases after sampling. Therefore, \(\Sigma\) is also \(\delta\)-sampled observable under the given conditions.

What do you think about this solution?

We value your feedback to improve our textbook solutions.

Access millions of textbook solutions in one place

  • Access over 3 million high quality textbook solutions
  • Access our popular flashcard, quiz, mock-exam and notes features
  • Access our smart AI features to upgrade your learning
Get Vaia Premium now
Access millions of textbook solutions in one place

Most popular questions from this chapter

Chapter 6

For linear systems, \((x, \sigma)\) and \((z, \sigma)\) are indistinguishable on \([\sigma, \tau]\) if and only if $$ \lambda_{x-z}^{\sigma, t}(\mathbf{0})=0 $$ for all \(t \in[\sigma, \tau]\); that is, \((x-z, \sigma)\) is indistinguishable from \((0, \sigma)\) using zero controls on \([\sigma, \tau) .\) Similarly, any two states \(x, z\) are indistinguishable if and only if \(x-z\) is indistinguishable from 0 .

Chapter 6

Show that, if \(m=p=1\) and if \(W_{\mathcal{A}}=P / q\) with \(P\) of degree $\leq n-1\( and \)q\( of degree \)n\(, then \)\mathcal{A}\( has rank \)n\( if and only if \)P$ and \(q\) are relatively prime. (Hint: Use Corollary \(6.6 .6\) and the fact that \(\mathcal{K}\left(\left(s^{-1}\right)\right)\) forms an integral domain.)

Chapter 6

If \((A, B, C)\) realizes \(\mathcal{A}\), then $\operatorname{rank} \mathcal{H}_{s, t}(\mathcal{A}) \leq \max \left\\{\operatorname{rank} \mathbf{O}_{s}(A, C), \operatorname{rank} \mathbf{R}_{t}(A, B)\right\\} \leq n\( for all \)s, t$.

Chapter 6

Let \(\omega_{i}, i=1, \ldots, k\) be \(k\) different positive real numbers. Show that there is some continuous-time time-invariant linear system with outputs and no inputs \(\Sigma=(A, 0, C)\) such that: \- \(\Sigma\) is observable, and \- for each set of \(2 k\) real numbers \(a_{i}, \varphi_{i}, i=1, \ldots, k\), there is some initial state \(x\) so that $$ \lambda_{x}^{0, t}=\eta(t)=\sum_{i=1}^{k} a_{i} \sin \left(2 \pi \omega_{i} t+\varphi_{i}\right) $$ for all \(t \geq 0\). Conclude from the above discussion that, if $$ \frac{1}{\delta}>2 \max _{i=1, \ldots, m}\left|\omega_{i}\right| $$ then the complete function \(\eta(t)\) can be recovered from the values $$ \eta(0), \eta(\delta), \eta(2 \delta), \ldots $$ for every set of \(a_{i}\) 's and \(\varphi_{i}\) 's.

Chapter 6

We consider again the parity check example discussed in Example 2.3.3. In particular, we shall see how to prove, using the above results, the last two claims in Exercise 2.3.4. The behavior to be realized is $\lambda(\tau, 0, \omega)=$ $$ \begin{cases}1 & \text { if } \omega(\tau-3)+\omega(\tau-2)+\omega(\tau-1) \text { is odd and } 3 \text { divides } \tau>0 \\ 0 & \text { otherwise }\end{cases} $$ and we take the system with $$ x:=\\{0,1,2\\} \times\\{0,1\\} $$ and transitions $$ \mathcal{P}((i, j), l):=(i+1 \bmod 3, j+l \bmod 2) $$ for \(i=1,2\) and $$ \mathcal{P}((0, j), l):=(1, l) \text {. } $$ The initial state is taken to be \((0,0)\), and the output map has \(h(i, j)=1\) if \(i=0\) and \(j=1\) and zero otherwise. (The interpretation is that \((k, 0)\) stands for the state " \(t\) is of the form \(3 s+k\) and the sum until now is even," while states of the type \((k, 1)\) correspond to odd sums.) This is clearly a realization, with 6 states. To prove that there is no possible (time-invariant, complete) realization with less states, it is sufficient to show that it is reachable and observable. Reachability follows from the fact that any state of the form \((0, j)\) can be obtained with an input sequence \(j 00\), while states of the type \((1, j)\) are reached from \(x^{0}\) using input \(j\) (of length one) and states \((2, j)\) using input \(j 0\). Observability can be shown through consideration of the following controls \(\omega_{i j}\), for each \((i, j):\) $$ \omega_{01}:=0, \omega_{00}:=100, \omega_{10}:=10, \omega_{11}:=00, \omega_{21}:=0, \omega_{20}:=0 . $$ Then, \(\omega_{01}\) separates \((0,1)\) from every other state, while for all other pairs \((i, j) \neq\) \((0,1)\), $$ \lambda_{(i, j)}\left(\omega_{\alpha \beta}\right)=1 $$ if and only if \((i, j)=(\alpha, \beta)\).

Join over 22 million students in learning with our Vaia App

The first learning app that truly has everything you need to ace your exams in one place.

  • Flashcards & Quizzes
  • AI Study Assistant
  • Smart Note-Taking
  • Mock-Exams
  • Study Planner
Join over 22 million students in learning with our Vaia App Join over 22 million students in learning with our Vaia App

Recommended explanations on Math Textbooks