Hybrid delayed/undelayed LQG compensation /
The problem of regulating the output of a stochastically perturbed finite dimensional linear time invariant continuous-time system by feedback control is considered. The regulation objectives are expressed in terms of a trade-off between steady-state control signal power and steady-state power of th...
| Main Author: | |
|---|---|
| Format: | Thesis eBook |
| Language: | English |
| Published: |
[Place of publication not identified] :
[publisher not identified] ;
2002.
|
| Subjects: | |
| Online Access: | Link to OAKTrust copy |
| Summary: | The problem of regulating the output of a stochastically perturbed finite dimensional linear time invariant continuous-time system by feedback control is considered. The regulation objectives are expressed in terms of a trade-off between steady-state control signal power and steady-state power of the system output. Sampled-data control with delay is taken as the benchmark control methodology. The three fundamental assumptions are: 1) this time delay is associated with the application of sampled-data control; 2) it imposes limits on the performance; 3) hardware costs associated with reducing the delay are unacceptable. As an alternative control methodology, the addition of delay-free feedback loops in parallel with the delayed sampled-data control loop are considered. The extra loop is constrained to be a proportional controller, with the idea that it might be cheaply implemented. It is hoped that limitations on performance imposed by the delay will be reduced while the flexibility associated with a digital implementation of sampled-data control is retained. This study proposes to use a comparison of the achievable trade-offs between control and output signal powers to compare the capabilities of the control structures. This is deemed a fair way to compare the structures with respect to the stated regulation objective. Controllers which achieve minimal performance (i.e. no other achievable performance has both smaller control power and smaller output power) are of interest. Standard discrete time LQG problems can be formulated to find these controllers. An example of evaluating the controller structures in terms of achievable regulator performance is carried out for an example plant with double integrator dynamics. As expected, simulation studies verify that the proposed control structures allow a smaller steady-state output power to be achieved with a given maximum steady-state control power. When more delay is present in the sampled-data control path, the improvement relative to standard sampled data control is larger. |
|---|---|
| Item Description: | "Major subject: Mechanical Engineering". Vita. |
| Physical Description: | ix, 60 leaves : illustrations ; 28 cm. Also available online. Issued also on microfiche from Lange Micrographics. |
| Bibliography: | Includes bibliographical references (leaves 58-59). |