

Type of Document Master's Thesis Author Fowler, Leslie Paige URN etd-5134151959631701 Title Application of the Filtered-X LMS Algorithm for Disturbance Rejection in Time-Periodic Systems Degree Master of Science Department Mechanical Engineering Advisory Committee
Advisor Name Title Harry H. Robertshaw Committee Chair Harley H. Cudney none William R. Saunders none Keywords
- adaptive control
- filtered-x LMS
- time-periodic
- disturbance rejection
- helicopter rotor blade
Date of Defense 1996-05-03 Availability unrestricted Abstract
Extensive disturbance rejection methods have
been established for time-invariant systems.
However, the development of these
techniques has not focused on application to
time-periodic systems in particular until
recently. The filtered-X LMS algorithm is
regarded as the best disturbance rejection
technique for aperiodic systems by many, as
has been proven in the acoustics industry for
rejecting unwanted noise. Since this is
essentially a feedforward approach, we might
expect its performance to be good with
respect to time-periodic systems in which the
disturbance frequency is already known. The
work presented in this thesis is an
investigation of the performance of the
filtered-X LMS algorithm for disturbance
rejection in time-periodic systems. Two cases
are examined: a generalized linear,
time-periodic system and the helicopter rotor
blade in forward flight. Results for the
generalized system show that the filtered-X
LMS algorithm does converge for
time-periodic disturbance inputs and can
produce very small errors. For the helicopter
rotor blade system the algorithm is shown to
produce very small errors, with a 96%, or 14
dB, reduction in error from the open-loop
system. The filtered-X LMS disturbance
rejection technique is shown to provide a
successful means of rejecting time-periodic
disturbances for time-periodic systems.
Files
Filename Size Approximate Download Time (Hours:Minutes:Seconds)
28.8 Modem 56K Modem ISDN (64 Kb) ISDN (128 Kb) Higher-speed Access etd.pdf 413.29 Kb 00:01:54 00:00:59 00:00:51 00:00:25 00:00:02
If you have questions or technical problems, please Contact DLA.