Abstract

This thesis deals with the experimental application of a system identification tech nique called pseudo-linear identification (PLID). PLID is a discrete-time, multi-input, multi-output (MEMO), state space, simultaneous parameter estimator and one step ahead state predictor of linear time invariant systems. No measurements are assumed perfect under PLED; that is the inputs and outputs are allowed to have zero mean white gaussian (ZMWG) additive noise. Furthermore, the states are also assumed to have additive ZMWG noise. Like most system identification techniques, PLED requires the system to be completely controllable and observable under the given actuator and sensor suite. The only firm assumption made on model structure is that the transfer function be strictly proper; that is, the frequency response is bounded and tends towards zero as frequency is in creased to infinity. Pole and zero locations are not confined; indeed, unstable systems can be identified, and furthermore, they can be controlled because PLED provides simultaneous one step ahead state predictions. Developed by Hopkins et. al. in 1988 [1], this method has seen little application (due in part to its youth); however, it is shown in the following pages to be a powerful technique for performing state space system identification, as well as on-line model order reduction. The experiment involves applying PLED to a 3 -Dimensional (3-D) kinematic truss structure (referred to here forward as the "testbed") in a batch mode (off-line). Batch mode identification, by definition, implies that the testbed does not change appreciably between the time it was identified and the time it will be controlled. For most kinematic structures, this is true. PLED can be used for real-time (on-line) system identification. However, due to the complexity of typical structures (e.g., flexible mechanical systems), and the high bandwidth of control (hundreds of hertz), this is not possible with current personal computer (PC) based controllers. Ultimately, the state space model generated by PLED will be used to design a closed loop controller for the testbed that will increase its damping twenty fold, from approximately 0.25% zeta to 5% zeta. Due to time constraints, we will only show simulation results of the closed loop system.

Library of Congress Subject Headings

System analysis; Stochastic systems; Algorithms

Publication Date

10-1-1997

Document Type

Thesis

Department, Program, or Center

Electrical Engineering (KGCOE)

Advisor

Hopkins, Mark

Advisor/Committee Member

Kempski, Mark

Advisor/Committee Member

Mathew, Athimoottil

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QA402 .V35 1997

Campus

RIT – Main Campus

Download

Share

COinS