Two true facts about linear systems, starting with the SISO (Single Input, Single Output) case.
Fundamental Solutions (SISO)
- The output of a linear system can always be written as a free response plus a forced response:
- There exists a set of fundamental solutions such that all solutions for can be computed from these
- The free response is a combination of them (weights set by the initial conditions)
- The impulse response is built from them as well
- This is the scalar preview of the fundamental matrix and general solution developed for MIMO/LTV systems
Corrected from the notes
- The quick notes wrote the convolution as and the kernel as ; since the free response starts at and the kernel depends on both and , they are written here as and
Modes and Poles (LTI)
- Fact: for LTI systems specifically, the fundamental solutions are complex exponentials:
- The are the modes of the system
- The are the poles of the system
- Generically, splitting a mode into its real and imaginary parts:
- with and
- sets the growth/decay envelope, sets the oscillation frequency
Corrected from the notes
- The quick notes wrote , the expansion’s last line as , and ; these are written here as , , and
Stability of a Fundamental Solution
- Fact: a fundamental solution (for a linear system) is stable if
- For LTI, , so
- and for all
- Thus
LTI Mode Stability
- A mode is stable exactly when its pole lies in the open left half of the complex plane
- The general (MIMO, time-varying) notion of stability is developed in Internal Stability
Speed of Convergence
- How fast does a stable mode die out? Use the 2% criterion (settling time ):
- Settling time is inversely proportional to : poles further left decay faster
Damping Ratio
- The damping ratio of a pole compares its decay rate to its distance from the origin:
- Equivalently:
- as (lightly damped, oscillation dominates)
- as (real pole, no oscillation)
- The MIMO version of “the response is a sum of modes” is the modal decomposition, where the poles become the eigenvalues of $A$