Projects proposed by industry
Dr.Owens is happy to consider any project proposed by
a company with which a sandwich student spent the summer.
Eigenstructure assignment by multivariable output feedback
A new algorithm for pole placement by multivariable output
feedback has recently been developed at Brunel. This project would entail
investigating the numerical properties of the new algorithm by applying
it to an test problem from the aerospace industry.
Optimal multivariable state feedback
A new algorithm for eigenstructure assignment by multi-input
state feedback has recently been developed at Brunel. The algorithm parameterizes
the design freedom available beyond eigenvalue assignment. This project
would entail investigating the use of the MATLAB optimisation toolbox to
determine an optimal controller that would assign a given
set of desired eigenvalues to a closed-loop system.
Reduced order state feedback through a reduced order
observer
Algorithms recently developed at Brunel allow for closed-loop
control by reduced order state feedback through a reduced order observer
in certain specified cases. This project would entail doing simulations
to establish the potential practical application of such a design strategy.
Eigenstructure assignment in linear descriptor systems
by state feedback
Algorithms have been developed at Brunel for eigenstructure
assignment in linear descriptor systems by state feedback. This project
would entail investigating the numerical properties of the new algorithms.
Computer control of a ball and beam apparatus
This project would involve implementing a digital controller
for a ball and beam apparatus in the Control Engineering laboratory using
a PC with A/D, D/A interfaces.
Modelling and control of ball and hoop apparatus
The system is nonlinear. There is a need for a lot of
careful measurement of system behaviour in order to model the apparatus.
Generating Describing Functions in SIMULINK
The describing function of a system with a hard nonlinearity
indicates whether the system will exhibit continuous oscillation at the
output. Traditionally the describing functions of standard non-linearities
were obtained from tables of describing functions. However, these tables
deal with a limited range of non-linearities. The describing functions
for nonlinearities not found in such tables may be generated using matlab
with simulink. In this project the student will investigate controller
design based on describing functions generated using matlab with simulink.
Generating Phase portraits in SIMULINK
The phase portrait of a second order system with a hard
nonlinearity gives an over view of the step responses of the system. Note
that the principle of superposition does not apply to a non-linear system
so the shape of the step response of the system depends on the magnitude
of the step input. The generation of phase portraits by hand is a time
consuming exercise however phase portraits may readily be generated using
the simulation package matlab with simulink. Historically phase portraits
have been used to design non-linear systems. In this project the student
would survey design methods
based on phase portraits and attempt to implement them
using simulink.