Tutorial 6

Discrete System and Delay-aware Controller Design

Dynamic Systems in Real World and CPS Domain
Pacific equatorial Jets
Any System that Changes w.r.t time is a Dynamic System.
We express them by the change of system variables w.r.t. time
i.e. dx/dt = f(x(t))

Change in the number of Vehicles

Change in the water flow at a traffic signal checkpoint
Horseshoes in hurricanes, 2010, Philip C Du’ Toit et al.

Harrier AV-8B military aircraft A Moving Vehicle / Traffic Flow in a Road

Change in the
Speed, Position of a
Vehicle

Change in the Speed,

Altitude of Jumper Jet
Feedback Systems, Karl J Astrom, R Murray
Automatic Cruise Control (ACC) System
• The purpose of a cruise control system is to maintain certain reference speed
despite external disturbances, such as changes in road traffic or road grade.
• This is accomplished by measuring the current vehicle speed and comparing
it to the desired or reference speed.
• This is also how a human driver drives, but there can be disturbances that can
lead to an unsafe situations (crashes).
• Automatic cruise control (ACC) intervenes and automatically adjusts the
throttle (and brakes) according to the implemented control law by observing
the current vehicle speed.
Physical Setup of ACC
• We simplify the model by assuming that we can control the force generated at road-tire surface directly and will
neglect the dynamics of the powertrain, tires, etc., that go into generating the force.
• A control force 𝑢 is applied on the vehicle, of mass 𝑚.
• The resistive forces, 𝑏𝑣, due to rolling resistance and wind drag, are assumed to vary linearly with the vehicle
velocity,𝑣.

• Summing forces in the x-direction and applying Newton's

2nd law, we arrive at the following system equation :
𝑚𝑣ሶ + 𝑏𝑣 = 𝑢
• The output equation is : 𝑦=𝑣
• We can consider 𝑣 as the changing state for this system
• Our Control Design Objective Should be reaching a
desired reference velocity 𝑣
State Space Modeling
• Usual Dynamic System is expressed using the following equation,
where 𝑓 is transfer function and 𝑔 is input to state transition function : 𝑥ሶ = 𝑓 (𝑥) + 𝑔(𝑥) ∗ 𝑢(𝑥, 𝑡)
• We can also Express it as : 𝑥ሶ = 𝐴(𝑥) + 𝐵(𝑥) ∗ 𝑢(𝑥, 𝑡)
• For linear time-invariant systems it can be expressed in State-Space Format
where we have A as state-transition matrix and B as input-to-state transition matrix : 𝑥ሶ = 𝐴𝑥 + 𝐵𝑢
[We can also linearize any polynomial transfer function at a certain expansion point, More on this later...]

State Space Model : 𝑥ሶ = 𝐴𝑥 + 𝐵𝑢;

Output: 𝑦 = 𝐶𝑥 + 𝐷𝑢;

−𝑏 1
𝑥ሶ = 𝑣ሶ = 𝑣 + 𝑢
𝑚 𝑚

𝑦= 1 𝑣

−𝑏 1
𝐴= ;𝐵 = ; 𝐶 = 1 ; 𝐷 = [0]
𝑚 𝑚
DC Motor Speed: System Modelling
R  armature resistance
L  armature inductance
K  electromotive force constant
J  moment of inertia
b  friction co-efficient
i  current
V  voltage is the input and
Ɵሶ  rotational speed of the shaft is the output

Following Newton's 2nd law and Kirchhoff's voltage law, we get

𝐽𝜃ሷ + 𝑏𝜃ሶ = 𝐾𝑖
𝑑𝑖
𝐿 + 𝑅𝑖 = 𝑉 − 𝐾 𝜃ሶ
𝑑𝑡

Q1. Model the State-space equations for the DC Motor Speed Control System?
DC Motor Speed: System Modelling
State space model format:
𝑥ሶ = 𝐴𝑥 + 𝐵𝑢
𝑦 = 𝐶𝑥 + 𝐷𝑢

x=  system state vector

y  output vector
u  input to the system

Q1. Model the State-space equations for the DC Motor Speed Control System
V  voltage is the input and
Sol. Ɵሶ  rotational speed of the shaft is the output
𝐽𝜃ሷ + 𝑏𝜃ሶ = 𝐾𝑖
𝑏ሶ 𝐾 𝑏 𝐾
⇒ 𝜃 = − 𝜃ሶ + 𝑖
ሷ 𝑑 𝜃ሶ − 0
𝐽 𝐽 𝐽 𝐽 𝜃ሶ
= + 1 𝑉
𝑑𝑖 𝑑𝑡 𝑖 𝐾 𝑅 𝑖
𝐿 + 𝑅𝑖 = 𝑉 − 𝐾 𝜃ሶ − − 𝐿
𝑑𝑡 𝐿 𝐿
𝑑𝑖 −𝐾 𝑅 1 𝜃ሶ
⇒ = 𝜃ሶ − 𝑖 + 𝑣 𝑦 = 10
𝑑𝑡 𝐿 𝐿 𝐿 𝑖
DC Motor Speed: System Modelling
V  voltage is the input and
Ɵሶ  rotational speed of the shaft is the output

𝑏 𝐾
𝑑 𝜃ሶ − 0
𝐽 𝐽 𝜃ሶ
= + 1 𝑉
𝑑𝑡 𝑖 𝐾 𝑅 𝑖
− − 𝐿
𝐿 𝐿
𝜃ሶ
𝑦 = 10
𝑖
Q2. Consider the following values of the constants and comment on the stability of the continuous-time system
J = 0.01;
b = 0.1;
K = 0.01;
R = 1;
L = 0.5;
DC Motor Speed: System Modelling
V  voltage is the input and
Ɵሶ  rotational speed of the shaft is the output

𝑏 𝐾
𝑑 𝜃ሶ − 0
𝐽 𝐽 𝜃ሶ
= + 1 𝑉
𝑑𝑡 𝑖 𝐾 𝑅 𝑖
− − 𝐿
𝐿 𝐿
𝜃ሶ
𝑦 = 10
𝑖
Q2. Consider the following values of the constants and comment on the stability of the continuous-time system
J = 0.01; b = 0.1; K = 0.01; R = 1; L = 0.5;

Sol. Real part of the poles must on left-side for the system to be stable. Poles are nothing but the eigen values of the
system matrix A. Using the above constants, we get,
𝑏 𝐾
−
𝐽 𝐽 −10 1
𝐴= =
𝐾 𝑅 −0.02 −2
− −
𝐿 𝐿
And, Eigen values of A are -10, -2. Therefore, the system is stable.
DC Motor Speed: System Modelling
V  voltage is the input and
Ɵሶ  rotational speed of the shaft is the output

𝑏 𝐾
𝑑 𝜃ሶ − 0
𝐽 𝐽 𝜃ሶ
= + 1 𝑉
𝑑𝑡 𝑖 𝐾 𝑅 𝑖
− − 𝐿
𝐿 𝐿
𝜃ሶ
𝑦 = 10
𝑖
Q3. Consider the following values of the constants and implement a pole-placement based controller using
Ackerman’s Method [poles at -12].
DC Motor Speed: System Modelling
Q3. Consider the following values of the constants and implement a pole-placement based controller using
Ackerman’s Method [poles at -12].

Sol. Using the above constants, we get,

𝑏 𝐾
−
𝐽 𝐽 −10 1 0
𝐴= = ,𝐵 =
𝐾 𝑅 −0.02 −2 2
− −
𝐿 𝐿
Step 1: Check the state controllability of the system
0 2
𝐶𝑀 = 𝐵 𝐴𝐵 𝐴2 𝐵 … 𝐴𝑛−1 𝐵 = 𝐵 𝐴𝐵 =
2 −4
Rank of CM is 2. Therefore, the system is controllable
Step 2: Compute K following
−1
𝐾 = 0 0 0 … 1 𝐵 𝐴𝐵 𝐴2 𝐵 … 𝐴𝑛−1 𝐵 𝜙 𝐴
𝜙 𝐴 = 𝐴𝑛 + 𝛼1 𝐴𝑛−1 + … + 𝛼𝑛−1 𝐴 + 𝛼𝑛 𝐼
Solve 𝑠 − 𝜇1 𝑠 − 𝜇2 = 0 to get 𝛼1 and 𝛼2 where 𝜇1 = 𝜇2 = −12
3.98 12
𝜙 𝐴 = 𝐴2 + 2𝐴 + 2𝐼 =
−0.24 99.98
1 0.5
𝐶𝑀1 =
0.5 0
𝐾 = [1.99 6]
DC Motor Speed: System Modelling
Q4. Discretize the system with sampling period = 0.6 sec
DC Motor Speed: System Modelling
Q4. Discretize the system with sampling period = 0.6 sec

Sol.
2 2
𝐴 ℎ
𝐴𝑑 = 𝑒 𝐴×ℎ = 𝐼 + 𝐴ℎ + + … ≈ 𝐼 + 𝐴ℎ
2!
ℎ ℎ 2𝑠2
𝐴
𝐵𝑑 = න 𝑒 𝐴×𝑠 𝐵 𝑑𝑠 = න 𝐵(𝐼 + 𝐴𝑠 + + … ) 𝑑𝑠 ≈ 𝐵ℎ
0 0 2!

Continuous-time system matrices:

𝑏 𝐾
−
𝐽 𝐽 −10 1 0
𝐴= = ,𝐵 =
𝐾 𝑅 −0.02 −2 2
− −
𝐿 𝐿
DC Motor Speed: System Modelling
Q4. Discretize the system with sampling period = 0.6 sec

Sol.
2 2
𝐴 ℎ
𝐴𝑑 = 𝑒 𝐴×ℎ = 𝐼 + 𝐴ℎ + + … ≈ 𝐼 + 𝐴ℎ
2!
ℎ ℎ 2𝑠2
𝐴
𝐵𝑑 = න 𝑒 𝐴×𝑠 𝐵 𝑑𝑠 = න 𝐵(𝐼 + 𝐴𝑠 + + … ) 𝑑𝑠 ≈ 𝐵ℎ
0 0 2!

Continuous-time system matrices:

𝑏 𝐾
−
𝐽 𝐽 −10 1 0
𝐴= = ,𝐵 =
𝐾 𝑅 −0.02 −2 2
− −
𝐿 𝐿

Discrete-time system matrices:

−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
Step 1: Controllability check
𝛾 = Γ 𝜙Γ 𝜙 2 Γ … 𝜙 𝑛−1 Γ
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
Step 1: Controllability check
𝛾 = Γ 𝜙Γ 𝜙 2 Γ … 𝜙 𝑛−1 Γ
0 0.72
= Γ 𝜙Γ =
1.2 −0.24
Rank(𝛾) = 2 The discrete system is controllable
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
Step 1: Controllability check
𝛾 = Γ 𝜙Γ 𝜙 2 Γ … 𝜙 𝑛−1 Γ
0 0.72
= Γ 𝜙Γ =
1.2 −0.24
Rank(𝛾) = 2 The discrete system is controllable
Step 2: Apply Ackermann’s formula
𝐾 = − 0 0 0 … 1 𝛾 −1 𝐻(𝜙)
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
Step 1: Controllability check
𝛾 = Γ 𝜙Γ 𝜙 2 Γ … 𝜙 𝑛−1 Γ
0 0.72
= Γ 𝜙Γ =
1.2 −0.24
Rank(𝛾) = 2 The discrete system is controllable
Step 2: Apply Ackermann’s formula
𝐾 = − 0 0 0 … 1 𝛾 −1 𝐻(𝜙)
𝐻 𝜙 = 𝜙 − 𝛼1 𝐼 𝜙 − 𝛼2 𝐼 … (𝜙 − 𝛼𝑛 𝐼) where 𝛼1 , 𝛼2 , … , 𝛼𝑛 are closed loop poles
DC Motor Speed: System Modelling
Q5. Implement a pole-placement based controller using Ackerman’s Method [poles at 0.2]
Sol.
Discrete-time system matrices:
−5 0.6 0
𝜙= Γ=
−0.012 −0.2 1.2
Step 1: Controllability check
𝛾 = Γ 𝜙Γ 𝜙 2 Γ … 𝜙 𝑛−1 Γ
0 0.72
= Γ 𝜙Γ =
1.2 −0.24
Rank(𝛾) = 2 The discrete system is controllable
Step 2: Apply Ackermann’s formula
𝐾 = − 0 0 0 … 1 𝛾 −1 𝐻(𝜙)
𝐻 𝜙 = 𝜙 − 𝛼1 𝐼 𝜙 − 𝛼2 𝐼 … (𝜙 − 𝛼𝑛 𝐼) where 𝛼1 , 𝛼2 , … , 𝛼𝑛 are closed loop poles
27.0328 -3.36
𝐻 𝜙 = 𝜙 − 0.2𝐼 𝜙 − 0.2𝐼 =
0.0672 0.1528
0.2778 0.8333
𝛾 −1 =
1.3889 0
𝐾 = [37.5456 − 4.6667]
DC Motor Speed: System Modelling
Q6. Derive a feed forward Gain
Sol.
DC Motor Speed: System Modelling
Q6. Derive a feed forward Gain
Sol.
1
𝐹 = 𝐶 𝐼 −𝜙−𝜏𝐾 −1𝜏 (note: C=[1 0])
DC Motor Speed: System Modelling
Q6. Derive a feed forward Gain
Sol.
1
𝐹 = 𝐶 𝐼 −𝜙−𝜏𝐾 −1𝜏 (note: C=[1 0])
𝐹 = 19.1311
DC Motor Speed: System Modelling
Q6. Derive a feed forward Gain
Sol.
1
𝐹 = 𝐶 𝐼 −𝜙−𝜏𝐾 −1𝜏 (note: C=[1 0])

𝐹 = 19.1311

Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.
DC Motor Speed: System Modelling
Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.
DC Motor Speed: System Modelling
Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.

𝜙 ≈ 𝐼 + 𝐴ℎ Γ1 𝐷𝑐 = 𝐷𝑐 𝐵 Γ0 = ℎ − 𝐷𝑐 𝐵
DC Motor Speed: System Modelling
Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.

𝜙 ≈ 𝐼 + 𝐴ℎ Γ1 𝐷𝑐 = 𝐷𝑐 𝐵 Γ0 = ℎ − 𝐷𝑐 𝐵

−5 0.6 0 0
𝜙𝑎𝑢𝑔 = −0.012 −0.2 0.6 Γ𝑎𝑢𝑔 = 0.6 𝐶𝑎𝑣𝑔 = [1 0 0]
0 0 0 1
DC Motor Speed: System Modelling
Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.

𝜙 ≈ 𝐼 + 𝐴ℎ Γ1 𝐷𝑐 = 𝐷𝑐 𝐵 Γ0 = ℎ − 𝐷𝑐 𝐵

−5 0.6 0 0
𝜙𝑎𝑢𝑔 = −0.012 −0.2 0.6 Γ𝑎𝑢𝑔 = 0.6 𝐶𝑎𝑣𝑔 = [1 0 0]
0 0 0 1

Q8. Derive a pole-placement based controller gain for the delayed system with poles at 0.2.
Follow the same steps and formula as discrete controller design only instead of 𝜙 and Γ, use 𝜙𝑎𝑢𝑔 and Γ𝑎𝑢𝑔 . Also, note
that, now the system’s dimension has increased by 1.
DC Motor Speed: System Modelling
Q7. Consider a delay of 0.3 sec and derive the state-space model of the delayed system
Sol.

𝜙 ≈ 𝐼 + 𝐴ℎ Γ1 𝐷𝑐 = 𝐷𝑐 𝐵 Γ0 = ℎ − 𝐷𝑐 𝐵

−5 0.6 0 0
𝜙𝑎𝑢𝑔 = −0.012 −0.2 0.6 Γ𝑎𝑢𝑔 = 0.6 𝐶𝑎𝑣𝑔 = [1 0 0]
0 0 0 1

Q8. Derive a pole-placement based controller gain for the delayed system with poles at 0.2.
Follow the same steps and formula as discrete controller design only instead of 𝜙 and Γ, use 𝜙𝑎𝑢𝑔 and Γ𝑎𝑢𝑔 . Also, note
that, now the system’s dimension has increased by 1.

Sol. Controllability matrix 𝛾𝑎𝑢𝑔 = [0 0.3600 -1.5120; 0.6000 0.4800 -0.1003; 1.0000 0 0]

𝐻(𝜙𝑎𝑢𝑔 ) = [27.0328 -3.3600 0.3600; 0.0672 0.1528 -0.3600; 0 0 0.0400]

−1 𝐻 𝜙
𝐾 = − 0 0 1 𝛾𝑎𝑢𝑔 𝑎𝑢𝑔 = [18.7800 -2.4184 0.4510]