The problem of jointly designing a robust controller and an intelligent
scheduler for networked control systems (NCSs) is addressed in this
thesis. NCSs composing of multiple plants that share a single channel
communication network with uncertain time-varying transmission times are
modeled as switched polytopic systems with additive norm-bounded
uncertainty. Switching is deployed to represent scheduling, the
polytopic uncertainty to overapproximatively describe the uncertain
time-varying transmission times. Based on the resulting NCS model and a
state feedback control law, the control and scheduling codesign problem
is then introduced and formulated as a robust (minimax) optimization
problem with the objective of minimizing the worst-case value of an
associated infinite time-horizon quadratic cost function. Five robust
codesign strategies are investigated for tackling the introduced
optimization problem, namely: Periodic control and scheduling (PCS),
Receding-horizon control and scheduling (RHCS), Implementation-aware
control and scheduling (IACS), Event-based control and scheduling
(EBCS), Prediction-based control and scheduling (PBCS). All these
codesign strategies are determined from LMI optimization problems based
on the Lyapunov theory. The properties of each are evaluated and
compared in terms of computational complexity and control performance
based on simulation and experimental study, showing their effectiveness
in improving the performance while utilizing the limited communication
resources very efficiently.
