Вы находитесь на странице: 1из 44

EE654A Project Presentation

Proof of Theorem 1 in “Switching Robust control for


Bilateral Teleoperation” by Cesar A. Lopez Martinez.

Group Members :
Vaibhav Sikerwar [17104082]
Satya Panuganti [14610]
Shashank Gupta [150661]
Theorem Statement
Theorem Statement
First Question
Construction of (*)
What is ?

Redheffer star product

It is used to combine two port systems In the above figure, the matrices for each
subsystem can be written as :.
What is ?
What is ?
Generalized plant model’s state space representation.

For a controller K, the generalized plant model’s state space model can be written as
follows :
Construction of (*)
Construction of (*)
Construction of (*)

(*) is then constructed in from the closed loop state representation obtained.
Pi & P p
Pi & P p

We use Pi to express the parametric uncertainty in the form of an Integral


Quadratic Constraint.
Pi & P p

We aim to minimize the L2 gain from input (the disturbances) to the output (the
constructed performance vector)

The above criterion can be expressed as a quadratic performance criterion :


Lyapunov Stability (A short recap)
Lyapunov Stability Theory

Let us consider time-invariant systems.


Lyapunov Stability Theory
Lyapunov Stability Theory
Lyapunov Stability Theory
Lyapunov Stability Theory
Stability of switched systems
Stability of switched Systems

Suppose that two individual subsystems are asymptotically stable, with trajectories as shown
on the left in the figure (the solid curve and the dotted curve). For different choices of the
switching signal, the switched system might be asymptotically stable or unstable (these two
possibilities are shown in the figure on the right).
Stability of switched Systems

Similarly, here two individual subsystems are not asymptotically stable, with trajectories as
shown on the left in the figure (the solid curve and the dotted curve). For different choices of
the switching signal, the switched system might be asymptotically stable or unstable (these
two possibilities are shown in the figure on the right).
Pointers

● It may be possible to stabilize a switched system by means of suitably constrained switching


even if all individual subsystems are unstable.
● Unconstrained switching may destabilize a switched system even if all individual subsystems
are stable.

There are two main problems in switching control :

● Finding conditions that guarantee asymptotic Stability of a switched system for


arbitrary switching signals.
● If a switched system is not asymptotically stable for arbitrary switching, identify
those switching signal for which it is asymptotically stable.

Theorem 1 of the paper focuses on the former problem.


KL class of functions
GUAS (Global Uniform Asymptotic Stability)
Common Lyapunov function

Theorem (2.1 in Liberzon)


If all systems in the family of switching controllers share a radially unbounded common
Lyapunov function, then the switched system is GUAS (Globally Uniform Asymptotic
Stability).
The above theorem can be worked out in a manner similar to the discussion in the
previous section for the Lyapunov stability theorem.
A (slightly) deeper journey into
the math
Parametric Uncertainties
A general class of uncertain continuous time systems

All the systems being discussed henceforth are in the above form.
Time varying parametric uncertainty

The above proposition provides a sufficient condition to conclude asymptotic stability of the time-varying
uncertainties in the system. Thus, the existence of K implies that arbitrary fast variation in the time-varying
parameter vector δ (·) may occur so as to guarantee asymptotic stability of the system.
It should be noted that if additional a priori information on the time-varying parameters in known, the above
proposition may lead to make the system too conservative.
Thus, a rate constraint on the switching may be added to reduce conservatism.
Time varying parametric uncertainty (Assumption)
Time varying parametric uncertainty
Time varying parametric uncertainty
Robust Performance
Robust Performance
Robust Performance
Finally, let us go back to the
original theorem which we had
set out to prove.
THEOREM 1!
THEOREM 1!
THEOREM 1!

Here, we are using the two IQCs to construct the supply function. Hence, we have shown
that the systems designed using theorem 1 are robustly dissipative against time-varying
uncertainties for the constructed supply function.
THEOREM 1!

Working out how the above result ensures


stability of the system
References

● Scherer, C., & Weiland, S. (2000). Linear matrix inequalities in control. Lecture
Notes, Dutch Institute for Systems and Control, Delft, The Netherlands, 3.
[Chapter 5]
● Lopez Martinez, C. A. (2015). A model-based robust control approach for bilateral
teleoperation systems Eindhoven: Technische Universiteit Eindhoven DOI:
10.6100/IR784500
[Chapters 2-3]
● Liberzon, D. (2012). Switching in systems and control. Springer Science & Business
Media.
[Chapter 2, Appendix A]

Вам также может понравиться