(Rhea HW3.A Brian Thomas)
 
(Updated by Brian Thomas -- Definition is correct; clarification added, with explanation)
 
(One intermediate revision by the same user not shown)
Line 1: Line 1:
'''Problem:''' 5 - Give a formal definition of a “stable system”. Give a formal definition of an unstable system.  
+
'''Problem:''' 5 - Give a formal definition of a “stable system”. Give a formal definition of an unstable system.
 
+
  
 
'''Solution:''' Consider system f, where input x(t) yields output y(t) = f(x(t)).  Consider bounded x(t), i.e. <math>\exists M_1 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |x(t)|\leq M_1</math>.  The system f can be considered '''stable''' iff <math>\forall x(t), \exists M_2 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |f(x(t))| \leq M_2</math>
 
'''Solution:''' Consider system f, where input x(t) yields output y(t) = f(x(t)).  Consider bounded x(t), i.e. <math>\exists M_1 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |x(t)|\leq M_1</math>.  The system f can be considered '''stable''' iff <math>\forall x(t), \exists M_2 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |f(x(t))| \leq M_2</math>
Line 10: Line 9:
  
 
(ie, The system f is stable iff there exists a bounded x(t) for which f(x(t)) is not bounded.)
 
(ie, The system f is stable iff there exists a bounded x(t) for which f(x(t)) is not bounded.)
 +
 +
 +
 +
For an LTI (Linear, Time-Invariant) system f, where <math>h(t) = f(\delta(t))</math>, the system f is considered '''stable''' iff <math>\int_{-\infty}^\infty |h[\tau ]| \,d \tau</math> is finite and '''unstable''' iff the above expression is infinite.
 +
 +
 +
 +
('''Note:''' The above is for continuous time (CT) systems.  Discrete time (DT) systems behave in almost exactly the same way; simply substitute all "(t)" with "[n]" (e.g., substitute "x(t)" with "x[n]", etc.) and integration with summation.)

Latest revision as of 16:03, 18 September 2008

Problem: 5 - Give a formal definition of a “stable system”. Give a formal definition of an unstable system.

Solution: Consider system f, where input x(t) yields output y(t) = f(x(t)). Consider bounded x(t), i.e. $ \exists M_1 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |x(t)|\leq M_1 $. The system f can be considered stable iff $ \forall x(t), \exists M_2 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |f(x(t))| \leq M_2 $

(ie, The system f is stable iff for all x(t), if x(t) is bounded, then f(x(t)) is bounded as well.)


Consider system f, where input x(t) yields output y(t) = f(x(t)). Consider bounded x(t), i.e. $ \exists M_1 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |x(t)|\leq M_1 $. The system f can be considered unstable iff $ \exists x(t) \text{ s.t. } \nexists M_2 \in \mathbb{R} \text{ s.t. } \forall t \in \mathbb{R}, |f(x(t))| \leq M_2 $

(ie, The system f is stable iff there exists a bounded x(t) for which f(x(t)) is not bounded.)


For an LTI (Linear, Time-Invariant) system f, where $ h(t) = f(\delta(t)) $, the system f is considered stable iff $ \int_{-\infty}^\infty |h[\tau ]| \,d \tau $ is finite and unstable iff the above expression is infinite.


(Note: The above is for continuous time (CT) systems. Discrete time (DT) systems behave in almost exactly the same way; simply substitute all "(t)" with "[n]" (e.g., substitute "x(t)" with "x[n]", etc.) and integration with summation.)

Alumni Liaison

Ph.D. 2007, working on developing cool imaging technologies for digital cameras, camera phones, and video surveillance cameras.

Buyue Zhang