Feedback Control System

Lesson 4:
Basic System Properties
Basic System Properties ▪ Any added/subtracted term in the
system relationship (except input
1. Linearity or output) must be constant or
• A system S is linear if it has the zero.
additivity property and the
homogeneity property. Let: 𝑦1 ≔ 3. Memory
𝑆𝑥1 and 𝑦2 ≔ 𝑆𝑥2 . (: = means “is • A system is memoryless (static) if
defined to be”) its output 𝑦 at time 𝑡 or 𝑛 depends
• Additivity:𝑦1 + 𝑦2 = 𝑆𝑥1 + 𝑥2 only on the input at that same time
• Homogeneity:𝑎𝑦1 = 𝑆𝑎𝑥1 , ∀𝑎 ∈ 𝐶 (present time).
• Homogeneity means that the • A system has memory (dynamic) if
response of 𝑆 to the scaled signal its output at time 𝑡 or 𝑛 depends on
𝑎𝑥1 is a times the response 𝑦1 = 𝑆𝑥1 . input values at some other times
• An important consequence is that (past or future time).
the response of a linear system to • Examples of Dynamic Memory
the 0 signal is the 0 signal. ▪ Time scaling
▪ Time shifting
2. Time-invariance ▪ Time operator
• A system S is time-invariant if its ▪ Integration based system
response to a time-shifted input
signal 𝑥[𝑛 − 𝑛𝑜 ] is equal to its 4. Causality
original response y[n] to x[n], but • A system is causal if its output at
also time shifted by no: y[n−no]. time 𝑡 or 𝑛 depends only on past or
• That is, if for y[n] ≔ Sx[n], y1[n] ≔ current values of the input
Sx[n−no] (independent of future values of
• The equality y1[n]=y[n−no] holds for input).
any integer no • If 𝑦1 ≔ 𝑆𝑥1 , 𝑦2 ≔ 𝑆𝑥2 , and
• Then the system is time-invariant. • 𝑥1 (𝜏) = 𝑥2 (𝜏), ∀𝜏𝜖(−∞, 𝑡] , then,
• Notes for Time-Invariant Systems: 𝑦1 (𝜏) = 𝑦2 (𝜏), ∀𝜏𝜖(−∞, 𝑡].
▪ No time scaling • This means that a causal system
▪ Coefficient should be constant subjected to two input signals that
coincide up to the current time 𝑡

1|Feedback Control
System
 produces outputs that also coincide Linearity: Linear
up to time 𝑡. 𝑣1 (𝑡) 𝑣2 (𝑡)
𝑖1 (𝑡) + 𝑖2 (𝑡) = +
• For noncausal systems, their 𝑅 𝑅
outputs up to time 𝑡 depends on 𝑣1 (𝑡) + 𝑣2 (𝑡)
𝑣1 (𝑡) + 𝑣2 (𝑡) → 𝑠𝑦𝑠𝑡𝑒𝑚 →
future values of the input signals, 𝑅
which may differ by assumption. (it Time-invariance: Time-invariant
can be dependent on present or 𝑣(𝑡 − 𝑎)
𝑖(𝑡 − 𝑎) =
past values) 𝑅
𝑣(𝑡 − 𝑎)
• Anti-causal system, dependent on 𝑣(𝑡 − 𝑎) → 𝑠𝑦𝑠𝑡𝑒𝑚 →
𝑅
future values on input only.
Memory: Static
𝑎𝑙𝑙 𝑡𝑖𝑚𝑒 𝑖𝑠 𝑎𝑡 𝑝𝑟𝑒𝑠𝑒𝑛𝑡
5. BIBO Stability
Causality: Causal
• A system 𝑆 is bounded-input
𝑎𝑙𝑙 𝑡𝑖𝑚𝑒 𝑖𝑠 𝑎𝑡 𝑝𝑟𝑒𝑠𝑒𝑛𝑡
bounded-output (BIBO) stable if for
BIBO Stability: Stable
any bounded input 𝑥, the
𝐴𝑡 𝑏𝑜𝑢𝑛𝑑𝑒𝑑 𝑣𝑜𝑙𝑡𝑎𝑔𝑒, 𝑐𝑢𝑟𝑟𝑒𝑛𝑡 𝑖𝑠 𝑎𝑙𝑠𝑜 𝑏𝑜𝑢𝑛𝑑𝑒𝑑
corresponding output 𝑦 is also Invertibility: Invertible
bounded. 𝐼𝑡 𝑖𝑠 𝑎 𝑜𝑛𝑒 − 𝑡𝑜 − 𝑜𝑛𝑒 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛
• Mathematically, the continuous-time
system 𝑦 𝑡 = 𝑆𝑥 𝑡 is BIBO stable if: Example:
∀𝐾1 > 0, ∃𝐾2 > 0 𝑠𝑢𝑐ℎ 𝑡ℎ𝑎𝑡 And inductor with resistance R and
𝑥(𝑡) < 𝐾1 , −∞ < 𝑡 < ∞ inductance L controls the input voltage v(t)
⇒ 𝑦(𝑡) < 𝐾2 , −∞ < 𝑡 < ∞ to its output current i(t)
𝑑𝑖
6. Invertibility 𝑣(𝑡) = 𝑅𝑖(𝑡) + 𝐿
𝑑𝑡
• A system 𝑆 is invertible if the input Answer: Linear, time-invariant, static,
signal can always be uniquely causal, BIBO stable, Invertible
recovered from the output signal.
• Mathematically, for 𝑥1 ≠ 𝑥2 , 𝑦1 = Example:
𝑆𝑥1 , 𝑦2 = 𝑆𝑥2 having 𝑦1 ≠ 𝑦2 . The An ideal operation amplifier integration
two distinct input signals cannot circuit has the following equation:
lead to the same output signal. 1 𝑡
𝑣𝑜𝑢𝑡 (𝑡) = ∫ 𝑣 (𝜏) 𝑑𝜏
• Functions are invertible if and only if 𝑅𝐶 −∞ 𝑖𝑛
they are both one-to-one (injective) Answer: Linear, time invariant, dynamic,
and onto (surjective). causal, BIBO stable, Invertible

Example: Example:
A resistor controls the input voltage v(t) to Determine the properties of the system:
its output current i(t) 𝑦[𝑛] = 𝑥[1 − 𝑛]
𝑣(𝑡) Answer: Linear, Time-variant, dynamic,
𝑖(𝑡) =
𝑅 non-causal, BIBO stable, Invertible

2|Feedback Control
System