RESEARCH PROJECTS
Synchronization of Lorenz System by Adaptive Observer
Suported partially
by Microtronic Inc (2000-2003)
Robust Control of Nonlinear System with Exogenous System
Fault-tolerant and Self-recovery Control for Nonlinear System
Near-optimal Control of Space Robotic System
Semiconductor Wafer Inspection System
Funded
by Microtronic Inc (2000-2003)
Synchronization of Lorenz System by Adaptive Observer
Chaotic communication has been an active research area during the
past decade. The research was initiated by Pecora and Carrol in 1990. They proposed a cascade scheme
that two identical chaotic systems can be synchronized, where one
system enslaves the other one. After this three modulation methods
have been considered: (1) Masking, which means add the key
(message signal) directly to the chaotic carrier. The carrier can be one state or combination of the
states of the transmitter; (2) Parameter modulation, which means
inject the message into the chaotic system to modulate the chaotic
carrier ; (3) Chaotic switching, which means a binary message is
transmitted by switching between two chaotic attractors generated
by two different parameter sets.
Application of using chaotic signals in secure telecommunication is
based on two considerations: 1) difficulty in extracting the
message signal by spectrum method since the chaotic carrier is
noise like broadband signal; 2)identical parameter set
between transmitter and receiver is unavailable to the third
party. Recent researcher shows that this kind of security
may be destroyed by some techniques other than by accessing the
identical parameter set. A nonlinear dynamic forecasting (NLD)
method has been suggested in By K. Short where an intercepted signal is used to extract the
information message by reconstructing the phase space of the
dynamic. This is done using a time-delay reconstruction. Another
method has been considered is to reconstruct partial state of the
chaotic dynamic using some return maps. The masked message can be
extracted by analyzing the attracting set of these return maps. Limitation of this algorithm is that it works well
only for frequency lower than the cutoff frequency. When the
frequency is around the cutoff frequency, the result is poor and
when the frequency is much higher than the cutoff frequency, the
algorithm doesn't work. In both of the algorithms mentioned above, parameters of the
system are still unknown at the receiver part. In application of
modulation scheme 2) and 3), assuming some binary signals are
modulated to constant parameters of chaotic system, for example,
during each modulation period, parameters of the chaotic system
are still constants. Thus, parameter estimation of the transmitter
is the key to solve this problem.
I proposed an adaptive observer with one state unavailable for Lorenz type system [ACC2003][ICNPAA]. System parameters are unknown but
assumed to be bounded and stay in a proper range that the Lorenz
system exhibits chaotic behavior. It's shown that synchronization
of states and estimation of the unknown system parameters have
been obtained. The proposed adaptation law guarantees asymptotic
convergence of the system. Simulation results also show the
effectiveness of the design and decoding of the message.
Robust Control of Nonlinear System with Exogenous System
A robust control is designed for a class of uncertain systems, and it is distinct and novel that the proposed control does not require any information of a bounding function on nonlinear uncertainties in the system. Instead, the uncertainties to be compensated for are generated by an exogenous system whose dynamics are either completed unknown or partially unknown. The only requirements on the exogenous system are that its unknown dynamics are bounded by a known function and that its output is bounded. The proposed robust control is based on a nonlinear observer that estimates the uncertainties [ACEXO][CDCEXO]. It is shown that, under different sets of conditions, local or semi-global or global stability of uniform ultimate boundedness or asymptotic stability can be achieved.
Fault-tolerant and Self-recovery Control for Nonlinear System
The problem of devising a fault-tolerant robust control for a class of nonlinear uncertain systems is investigated. Possible failures of the sensors measuring the state variables are considered, and robust measures are developed to identify the stability-vulnerable failures. Based on evaluation of the robust measures, a fault-tolerant robust control will switch itself among several robust control strategies designed under normal operation and under specific faulty conditions[AMFAU][CDCFAU][IASTED]. It is shown that, under mild conditions, the proposed scheme guarantees not only the desired performance under normal operations but also robust stability and best achievable performance when there is a sensor failure of any kind.
Near-optimal Control of Space Robotic
|
A robotic system considered in this paper consists of an n-DOF robotic manipulator mounted onto an attitude controlled base (which is either a space shuttle or a space station). To minimize control energy of thrust in maintaining the attitude of the base, nonlinear optimal control problem of robotic manipulator is formulated. Nonlinear optimal controls are outlined, based on which a new nonlinear near-optimal control is proposed for space robotic systems[JRAOPT][CCAOPT]. Compared to an optimal control which must be solved off-line and stored numerically, the proposed control scheme can easily be implemented real time, and its closeness to the optimal control can be measured. Stability results under the proposed sub-optimal control are obtained, and a simulation example is included to demonstrate its effectiveness. Semiconductor Wafer Inspection System
References
|
