ME-339: Control Engineering
Lec 02: Linearization
Instructor:
Dr. Khawaja Fahad Iqbal
Assistant Professor (Department of Robotics & AI, SMME, NUST)
Co-Principal Investigator (Intelligent Robotics Lab, NCAI)
04 Feb 2025 Department of Robotics and AI, SMME, NUST 1/31
LTI Systems
• The general LTI model is: A - Given by the Physics of
the system (we cannot
control this)
B – How input changes the
state (actuators)
C – How we get information
about the state (sensors)
• Main Question: How should the input be selected?
• But First: How can such systems be understood? And
where do they come from?
04 Feb 2025 Department of Robotics and AI, SMME, NUST 2/31
Example 1: The Car Model
• The car model for cruise control:
• Suppose we only care about / can measure the velocity:
• Suppose we care about / can measure the position:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 3/31
Example 2: Two Simple “Robots”
• Consider two “robots” on a line.
• The Rendezvous Problem: Have them meet at the same location.
• Idea: Have them aim towards each other:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 4/31
Example 3: Unicycle Robot
• We need to be more systematic/ clever when it comes to generating LTI
models from nonlinear systems!
04 Feb 2025 Department of Robotics and AI, SMME, NUST 5/31
Outline
Linearization
Stability
04 Feb 2025 Department of Robotics and AI, SMME, NUST 6/31
Outline
Linearization
Stability
04 Feb 2025 Department of Robotics and AI, SMME, NUST 7/31
Linearization
“Classifying systems as linear and non-linear is like
classifying objects in the Universe as bananas and
non-bananas.” -unknown
04 Feb 2025 Department of Robotics and AI, SMME, NUST 8/31
Linearization
• Given a non-linear model
• We want to find a “local”, linear model around an operating point
• The new equations of motion become
04 Feb 2025 Department of Robotics and AI, SMME, NUST 9/31
Linearization
04 Feb 2025 Department of Robotics and AI, SMME, NUST 10/31
Linearization
04 Feb 2025 Department of Robotics and AI, SMME, NUST 11/31
Computing the Jacobians
04 Feb 2025 Department of Robotics and AI, SMME, NUST 12/31
Computing the Jacobians
04 Feb 2025 Department of Robotics and AI, SMME, NUST 13/31
Computing the Jacobians
04 Feb 2025 Department of Robotics and AI, SMME, NUST 14/31
Example 1: Inverted Pendulum
04 Feb 2025 Department of Robotics and AI, SMME, NUST 15/31
Example 1: Inverted Pendulum
04 Feb 2025 Department of Robotics and AI, SMME, NUST 16/31
Example 1: Inverted Pendulum
04 Feb 2025 Department of Robotics and AI, SMME, NUST 17/31
Example 2: Unicycle
04 Feb 2025 Department of Robotics and AI, SMME, NUST 18/31
Conclusion
• Sometimes the linearization give reasonable models and sometimes
they do not.
• Despite the fact that they are only local approximations, they are
remarkably useful (when they work).
04 Feb 2025 Department of Robotics and AI, SMME, NUST 19/31
Outline
Control Theory
LTI Systems
Linearization
Stability
04 Feb 2025 Department of Robotics and AI, SMME, NUST 20/31
Stability of LTI System
• Let’s figure out how such systems behave.
• Start by ignoring the input term:
• What is the solution to this system?
04 Feb 2025 Department of Robotics and AI, SMME, NUST 21/31
Solving the ODE
• If everything is scalar:
• How do we know?
• For higher-order systems, we just get a matrix version of this:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 22/31
Stability
• First order of business is always trying to figure out
if the system “blows up” or not.
• Recall the control design objectives:
– Stability
– Tracking
– Robustness
– Other objectives
04 Feb 2025 Department of Robotics and AI, SMME, NUST 23/31
Scalar Systems
• We start with scalar systems to get some intuition about what is going on:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 24/31
Scalar Systems
• We start with scalar systems to get some intuition about what is going on:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 25/31
Scalar Systems
• We start with scalar systems to get some intuition about what is going on:
04 Feb 2025 Department of Robotics and AI, SMME, NUST 26/31
Three Cases
• Asymptotically Stable:
• Unstable:
• Critically Stable: in-between (doesn’t blow up but doesn’t go to zero either)
04 Feb 2025 Department of Robotics and AI, SMME, NUST 27/31
From Scalars to Matrices
• We cannot say that A>0, but we can do the next best
thing – eigenvalues!
• The eigenvalues tell us how the matrix A “acts” in
different directions (eigenvectors)
04 Feb 2025 Department of Robotics and AI, SMME, NUST 28/31
Stability
• Asymptotically Stable (if and only if):
• Unstable (if):
• Critically Stable (only if):
• Critically Stable (if): one eigenvalue is 0 and the rest have negative real part OR two
purely imaginary eigenvalues and the rest have negative real part
04 Feb 2025 Department of Robotics and AI, SMME, NUST 29/31
Thank you for your attention.
Any Questions?
04 Feb 2025 Department of Robotics and AI, SMME, NUST 30/31
