Professional Documents
Culture Documents
ecam.fr
1. INTRODUCTION
Continuous-Time System
representation and Modeling
𝑑1 … 𝑑𝑟
𝑢1 𝑦1
𝑢2
Inputs System Outputs
⋮
⋮
𝑢𝑚 𝑦𝑝
ecam.fr 3
Example: Representation of the Mass-Spring-Damper System
1
𝑘 𝐾
• Transfer Function 𝐺 𝑠 = =
𝑏 𝑚 2ζ 1
1 + 𝑠 + 𝑠2 1 + 𝜔 𝑠 + 𝜔 𝑠
𝑘 𝑘 0 0
0 1 𝑥 0
𝑥1 1
• State-Space representation = −𝑘 −𝑏 𝑥 + 1 𝑢
𝑥2 2
𝑚 𝑚 𝑚
𝒏
𝒚 + 𝒂𝟏 𝒚(𝒏−𝟏) + ⋯ + 𝒂𝒏−𝟏 𝒚 + 𝒂𝒏 𝒚 = 𝒃𝟎 𝒖 𝒎
+ 𝒃𝟏 𝒖 𝒎−𝟏
+ ⋯ + 𝒃𝒎−𝟏 𝒖 + 𝒃𝒎 𝒖 (1)
• If the initial conditions are equal to zero, this Equation (1) can be also written as
𝟏 × 𝒔𝒏
Here the coefficient that multiplies the term 𝑠 𝑛 is equal to 1
• Diagonal
canonical form
Controllable Observable
• Jordan
canonical form canonical form
canonical form
ecam.fr 8
2. CANONICAL FORMS
2.1. CONTROLLABLE CANONICAL FORM
2.1.1. CASE OF PROPER TRANSFER FUNCTION
• Consider the system which is described by the transfer function (3) (Case
of proper transfer function), it can be put in the state-space
representation which is called a controllable canonical form
𝑨𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆 𝑩𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
State Equation
(6)
Output Equation 𝑥1
𝑥2
𝑦 = 𝑏𝑛 − 𝑎𝑛 𝑏0 | 𝑏𝑛−1 − 𝑎𝑛−1 𝑏0 ⋯ ⋯| 𝑏1 − 𝑎1 𝑏0 ⋮ + 𝑏0 𝑢 (7)
𝑥𝑛
𝑪𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆 𝑫𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
ecam.fr 9
SEE THE DEMONSTRATION IN THE APPENDIX A
ecam.fr 10
• Using the state equation (6) as well as the output equation (7) (the
controllable canonical form ) then we can plot the block diagram
ecam.fr 11
2.1. CONTROLLABLE CANONICAL FORM
2.1.2. CASE OF STRICTLY PROPER TRANSFER FUNCTION
• Consider the system which is described by the transfer function (4) (Case
of strictly proper transfer function), it can be put in the state-space
representation which is called a controllable canonical form
𝒀(𝒔) 𝒃𝟎 𝒔𝒎 + 𝒃𝟏 𝒔𝒎−𝟏 + ⋯ + 𝒃𝒎−𝟏 𝒔 + 𝒃𝒎
=
𝑼(𝒔) 𝒔𝒏 + 𝒂𝟏 𝒔𝒏−𝟏 + ⋯ + 𝒂𝒏−𝟏 𝒔 + 𝒂𝒏
State Equation 𝒙𝟏 𝒙𝟏
𝟎 𝟏 ⋯ 𝟎 𝟎
𝒙𝟐 𝒙𝟐 𝟎
⋮ 𝟎 𝟏 𝟎 𝒙𝟑 + 𝟎 𝒖
⋮ =
⋮ ⋮ 𝟎 𝟏 ⋮
⋮ ⋮
−𝒂𝒏 −𝒂𝒏−𝟏 ⋯ −𝒂𝟏 𝒙
𝒙𝒏 𝒏 𝟏
Output Equation 𝒙𝟏
𝒙𝟐
𝒚 𝒕 = 𝒃𝒎 𝒃𝒎−𝟏 ⋯ 𝒃𝟏 𝒃𝟎 ⋮ + 𝟎 ×𝒖
⋮
𝒙𝒏
ecam.fr 12
SEE THE DEMONSTRATION IN THE APPENDIX B
ecam.fr 13
2.1. CONTROLLABLE CANONICAL FORM
2.1.3. CASE OF A TRANSFER FUNCTION WITH SIMPLE NUMERATOR
• Consider the system which is described by the transfer function (5) (Case
of a transfer function with simple numerator, it can be put in the state-
space representation which is called a controllable canonical form
𝒀(𝒔) 𝒃𝒎
𝑯 𝒔 = = 𝒏
𝑼(𝒔) 𝒔 + 𝒂𝟏 𝒔𝒏−𝟏 + ⋯ + 𝒂𝒏−𝟏 𝒔 + 𝒂𝒏
State Equation 𝒙𝟏 𝒙𝟏
𝟎 𝟏 ⋯ 𝟎 𝟎
𝒙𝟐 𝒙𝟐 𝟎
⋮ 𝟎 𝟏 𝟎 𝒙𝟑 + 𝟎 𝒖
⋮ =
⋮ ⋮ 𝟎 𝟏 ⋮
⋮ ⋮
−𝒂𝒏 −𝒂𝒏−𝟏 ⋯ −𝒂𝟏 𝒙
𝒙𝒏 𝒏 𝟏
Output Equation
𝒙𝟏
𝒙𝟐
𝒚 𝒕 = 𝒃𝒎 𝟎 ⋯ 𝟎 ⋮ + 𝟎𝒖
𝒙𝒏
ecam.fr 14
SEE THE DEMONSTRATION IN THE APPENDIX C
ecam.fr 15
2. CANONICAL FORMS
2.2. OBSERVABLE FORM
• Let ‘s consider the system described by the transfer function (3). It can be
put in the following state-space representation which is called an observable
canonical form (here only the case of proper TF was treated).
State Equation 𝑨𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 𝑩𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆
𝑥1 0 0 ⋯ 0 −𝑎𝑛 𝑥1 𝑏𝑛 − 𝑎𝑛 𝑏0
𝑥2 1 0 ⋯ 0 −𝑎𝑛−1 𝑥2 𝑏𝑛−1 − 𝑎𝑛−1 𝑏0 (8)
⋮ = 0 1 0 ⋮ ⋮ ⋮ + ⋮ 𝑢
⋮ ⋮ 0 ⋱ 0 −𝑎2 ⋮ ⋮
𝑥𝑛 0 ⋯ 0 1 −𝑎1 𝑥𝑛 𝑏1 − 𝑎1 𝑏0
𝑥1
Output Equation 𝑥2
(9)
𝑦 = 0 ⋯ ⋯ 0 1 ⋮ + 𝑏0 𝑢
⋮
𝑥𝑛
𝑪𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 𝑫𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆
ecam.fr 16
Remark 1
From the equations (8)-(9), we can deduce the following
𝑨𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 = 𝑨𝑻𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
𝑩𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 = 𝑪𝑻𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
𝑪𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 = 𝑩𝑻𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
𝑫𝒐𝒃𝒔𝒆𝒓𝒗𝒂𝒃𝒍𝒆 = 𝑫𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
• Here only the case of proper transfer function has been developed.
• Using the Remark 1, the other cases can be simply deduced from the controllable
Canonical form)
Control Theory 2B- SSR in Canonical
ecam.fr forms
17
2. CANONICAL FORMS
2.3. DIAGONAL FORM
𝑌(𝑠) 𝑐1 𝑐2 𝑐𝑛
= 𝑏0 + + + ⋯+ (10)
𝑈(𝑠) 𝑠 + 𝑝1 𝑠 + 𝑝2 𝑠 + 𝑝𝑛
• If 𝑚 = 𝑛, then 𝑏0 is constant
• If 𝑚 < 𝑛, then 𝑏0 = 0
ecam.fr 18
The diagonal canonical form of the state-space representation of this system is
given by
𝑥1 −𝑝1 0 ⋯ 0 𝑥1 1
𝑥2 0 −𝑝2 ⋯ 0 𝑥2 1 (11)
= ⋱ ⋮ + 𝑢
⋮ ⋮ ⋮ ⋮ ⋮
𝑥𝑛 0 0 ⋯ −𝑝𝑛 𝑥𝑛 1
𝑥1
⋯ 𝑐𝑛 𝑥2
𝑦 = 𝑐1 𝑐2
⋮ + 𝑏0 𝑢 (12)
𝑥𝑛
For this case, the preceding diagonal canonical form must be modified into
the Jordan canonical form
Suppose, for example, that the 𝒑𝒊 ′s are different from one another, except
that the first three 𝒑𝒊 ’s are equal, or 𝑝1 = 𝑝2 = 𝑝3 . Then the factored form of
𝑌(𝑠)
becomes
𝑈(𝑠)
𝑌(𝑠) 𝑐1 𝑐2 𝑐3 𝑐4 𝑐𝑛
= 𝑏0 + 3+ 𝑠+𝑝 2 + 𝑠 + 𝑝 + 𝑠 + 𝑝 + ⋯+ 𝑠 + 𝑝
𝑈(𝑠) 𝑠 + 𝑝1 1 1 4 𝑛
(13)
𝑥1
𝑥2 (14)
𝑦 = 𝑐1 𝑐2 ⋯ 𝑐𝑛
⋮ + 𝑏0 𝑢
𝑥𝑛
𝑌(𝑠) 𝑠+3
= 2
𝑈(𝑠) 𝑠 + 3𝑠 + 2
Controllable Canonical
Form
Observable Canonical
Form
In this section we will consider the transformation of the system model from
transfer function to state space, and vice versa
3.1. TRANSFORMATION FROM TRANSFER FUNCTION TO STATE SPACE
Let us write the closed-loop transfer function as
𝑨, 𝑩, 𝑪, 𝑫 = 𝒕𝒇𝟐𝒔𝒔(𝒏𝒖𝒎, 𝒅𝒆𝒏)
Remark
It is important to note that the state-space representation for any system is not
unique There are many state space representations for the same system. The
MATLAB command gives one possible state-space representation
(15)
There are many possible state-space representations for this system. One
possible state-space representation is
(16)
(17)
MATLAB transforms the transfer function given by Equation (15) into the
state-space representation given by Equations (16) and (17).
For the example considered here, MATLAB Program (see Figure 2) will
produce matrices A, B, C, and D
To obtain the transfer function from state-space equations, use the following
command
"𝒊𝒖" must be specified for systems with more than one input
For example, if the system has three inputs (𝒖𝟏 , 𝒖𝟐 , 𝒖𝟑 ), then iu must
be either 1, 2, or 3, where 1 implies 𝒖𝟏 , 2 implies 𝒖𝟐 , and 3 implies 𝒖𝟑
Consider a system with multiple inputs and multiple outputs. When the system
has more than one output, the command
produces transfer functions for all outputs to each input. (The numerator
coefficients are returned to matrix NUM with as many rows as there are outputs.)
Example
Consider the system defined by
ecam.fr 29
This system involves two inputs and two outputs: Four transfer functions
𝑌 (𝑠) 𝑌 (𝑠) 𝑌 (𝑠) 𝑌 (𝑠)
are involved: 𝑈1 (𝑠) , 𝑈21 (𝑠) , 𝑈2 (𝑠) , 𝑎𝑛𝑑 𝑈2 (𝑠)
1 1 2
(When considering input 𝒖𝟏 , we assume that input 𝒖𝟐 is zero and vice
versa)
ecam.fr 30
This is the MATLAB representation of
the following four transfer functions
Figure 3
Control Theory 2B- SSR in Canonical
ecam.fr forms
31
3.3. CORRELATION BETWEEN STATE-SPACE REPRESENTATION (SSR)
AND TRANSFER FUNCTION (TF)
𝑠𝑋 𝑠 =𝐴𝑋 𝑠 +𝐵𝑈 𝑠
𝑌 𝑠 = 𝐶 𝑋 𝑠 + 𝐷 𝑈(𝑠)
𝑠 𝐼𝑛×𝑛 − 𝐴 𝑋 𝑠 = 𝐵 𝑈 𝑠
ecam.fr 32
• In the output equation 𝑌 𝑠 , We replace 𝑋(𝑠) by its expression, then we get
𝑌 𝑠 =𝐶𝑿 𝒔 +𝐷𝑈 𝑠
𝑌 𝑠 = 𝐶 𝒔 𝑰 − 𝑨 −𝟏 𝑩 𝑼 𝒔 + 𝐷 𝑈 𝑠
𝑌 𝑠 = 𝐶 𝑠 𝐼 − 𝐴 −1 𝐵 + 𝐷 𝑈(𝑠)
• The correlation between the SSR and TF can be deduced using the following formula
𝒀(𝒔) −𝟏 𝑩 + 𝑫
=𝑪 𝒔𝑰−𝑨
𝑼(𝒔)
ecam.fr 33
APPENDIX A:
DEMONSTRATION OF THE CONTROLLABLE CANONICAL FORM
CASE OF PROPER TRANSFER FUNCTION
ecam.fr 34
Exercise (Demonstration of the controllable canonical form)
Consider the transfer function system defined by (case of proper TF)
ecam.fr 35
Solution
𝑏1 = 𝑏0 𝑎1 + 𝛼1 𝛼1 = 𝑏1 − 𝑏0 𝑎1
𝑏2 = 𝑏0 𝑎2 + 𝛼2 𝛼2 = 𝑏2 − 𝑏0 𝑎2
⋮ ⋮ (E.2)
⋮ ⋮
𝑏𝑛 = 𝑏0 𝑎𝑛 + 𝛼𝑛 𝛼𝑛 = 𝑏𝑛 − 𝑏0 𝑎𝑛
ecam.fr 36
we can substitute 𝛼1 , 𝛼2 , … , 𝛼𝑛 by their expressions (E.2) into the equation
(E.1), we get
𝒀𝟏 𝒔 𝑏1 − 𝑏0 𝑎1 𝑠 𝑛−1 + 𝑏2 − 𝑏0 𝑎2 𝑠 𝑛−2 + ⋯ + 𝑏𝑛 − 𝑏0 𝑎𝑛
= (E.6)
𝑈(𝑠) 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + ⋯ +𝑎𝑛
ecam.fr 37
We can write the following
𝑌1 (𝑠) 𝑌1 (𝑠) 𝑊(𝑠)
= ×
𝑈(𝑠) 𝑊(𝑠) 𝑈(𝑠)
𝑊(𝑠)
Where is defined as follows
𝑈(𝑠)
𝑾 𝒔 1
= 𝑛
𝑈(𝑠) 𝑠 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛 (E.7)
𝑌1 (𝑠)
And is defined as follows
𝑊(𝑠)
𝑌1 (𝑠)
= 𝑏1 − 𝑏0 𝑎1 𝑠 𝑛−1 + 𝑏2 − 𝑏0 𝑎2 𝑠 𝑛−2 + ⋯ 𝑏𝑛 − 𝑏0 𝑎𝑛
𝑊(𝑠) (E.8)
ecam.fr 38
• From (E.7) we have
𝑠 𝑛 𝑊 𝑠 + 𝑎1 𝑠 𝑛−1 𝑊 𝑠 + ⋯ 𝑎𝑛 𝑊 𝑠 = 𝑈(𝑠)
• Applying the Laplace inverse with initial conditions equal to zeros we get
𝑤 𝑤
𝑤 𝑤
𝑋= 𝑤 𝑋= 𝑤
⋮ ⋮
𝑤 (𝑛−1) 𝑤 (𝑛)
ecam.fr 39
Using (E.9) we can deduce the following
𝒏 𝒏−𝟏
𝒘 𝒕 = −𝒂𝟏 𝒘 𝒕 − ⋯ − 𝒂𝒏 𝒘 𝒕 + 𝒖(𝒕)
𝑤 𝟎 𝟏 𝟎⋯ 𝟎 𝑤 𝟎
𝑤 𝟎 𝟎 𝟏⋯ 𝟎 𝑤 𝟎
𝑤 = ⋮ ⋮ 𝟎⋱ ⋮ 𝑤 + ⋮ 𝑢
⋮ ⋮ ⋮ ⋮ 𝟎 𝟏 ⋮ 𝟎
𝑤 (𝑛) −𝒂𝒏 −𝒂𝒏−𝟏 ⋯ … −𝒂𝟏 𝑤 (𝑛−1) 𝟏
𝑩𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
𝑨𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
• Applying the Laplace inverse to (E.5) and then Substituting it into (E.10)
we obtain
𝑦 𝑡 − 𝑏0 𝑢(𝑡) = 𝑏1 − 𝑏0 𝑎1 𝑤 𝑛−1 𝑡 + 𝑏2 − 𝑏0 𝑎2 𝑤 𝑛−2 𝑡 + ⋯ 𝑏𝑛 − 𝑏0 𝑎𝑛 𝑤(𝑡)
ecam.fr 40
Which can be also written as follows
𝒚 𝒕 = 𝒃𝟎 𝒖 𝒕 + 𝒃𝟏 − 𝒃𝟎 𝒂𝟏 𝒘 𝒏−𝟏 𝒕 + 𝒃𝟐 − 𝒃𝟎 𝒂𝟐 𝒘 𝒏−𝟐
𝒕 + ⋯ 𝒃𝒏 − 𝒃𝟎 𝒂𝒏 𝒘(𝒕)
𝑤
𝑤
𝑦 𝑡 = 𝒃𝒏 − 𝒃𝟎 𝒂𝒏 𝒃𝒏−𝟏 − 𝒃𝟎 𝒂𝒏−𝟏 ⋯⋯ 𝒃𝟏 − 𝒃𝟎 𝒂𝟏 𝑤 + 𝒃𝟎 𝑢(𝑡)
⋮
𝑤 (𝑛−1)
𝑫𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
𝑪𝒄𝒐𝒏𝒕𝒓𝒐𝒍𝒍𝒂𝒃𝒍𝒆
ecam.fr 41
APPENDIX B:
DEMONSTRATION OF THE CONTROLLABLE CANONICAL FORM
CASE OF STRICTLY PROPER TRANSFER FUNCTION
ecam.fr 42
Let’s consider the system
𝑾 𝒔 1
= (A.2)
𝑈(𝑠) 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛
𝑌(𝑠)
And is defined as follows
𝑊(𝑠)
𝑌(𝑠) (A.3)
= 𝑏0 𝑠 𝑚 + 𝑏1 𝑠 𝑚−1 + ⋯ + 𝑏𝑚−1 𝑠 + 𝑏𝑚
𝑊(𝑠)
ecam.fr 43
• From (A.2) we have
𝑠 𝑛 𝑊 𝑠 + 𝑎1 𝑠 𝑛−1 𝑊 𝑠 + ⋯ 𝑎𝑛 𝑊 𝑠 = 𝑈(𝑠)
• Applying the Laplace inverse with initial conditions equal to zeros we get
𝑤 𝑤
𝑤 𝑤
𝑋= 𝑤 𝑋= 𝑤 (A.5)
⋮ ⋮
𝑤 (𝑛−1) 𝑤 (𝑛)
ecam.fr 44
Using (A.4) and (A.5) then we can write
𝒘 ⋯ 𝒘 𝟎
𝟎 𝟏 𝟎
𝒘 𝟏 𝒘 𝟎
⋮ 𝟎 𝟎 𝒘
𝒘 = + 𝟎 𝒖
⋮ ⋮ 𝟎 𝟏 (A.7)
⋮ ⋮ ⋮
−𝒂𝒏 −𝒂𝒏−𝟏 ⋯ −𝒂𝟏
𝒘(𝒏) 𝒘(𝒏−𝟏) 𝟏
Using (A.3) we can write
𝑌 𝑠 = 𝑏0 𝑠 𝑚 𝑊 𝑠 + 𝑏1 𝑠 𝑚−1 𝑊 𝑠 + ⋯ + 𝑏𝑚−1 𝑠 𝑊 𝑠 + 𝑏𝑚 𝑊(𝑠) (A.8)
Case 1: If 𝑚 = 𝑛 − 1 (we treat here the stricly propoer TF, then the max that 𝑚
can take is 𝑛 − 1 then (A.8) will be equal to
• Applying the Laplace inverse to (A.8) with initial conditions equal to zeros we get
𝑑𝑤 (𝑚) (𝑡) 𝑑𝑤 (𝑚−1) (𝑡) 𝑑𝑤 𝑡
𝑦 𝑡 = 𝑏0 + 𝑏1 + ⋯ + 𝑏𝑚−1 + 𝑏𝑚 𝑤(𝑡) (A.9)
𝑑𝑡 𝑚 𝑑𝑡 𝑚−1 𝑑𝑡
ecam.fr 45
𝑑𝑤 (𝑚) (𝑡) 𝑑𝑤 (𝑚−1) (𝑡) 𝑑𝑤 𝑡
𝑦 𝑡 = 𝑏0 + 𝑏1 + ⋯ + 𝑏𝑚−1 + 𝑏𝑚 𝑤(𝑡)
𝑑𝑡 𝑚 𝑑𝑡 𝑚−1 𝑑𝑡
ecam.fr 46
APPENDIX C:
DEMONSTRATION OF THE CONTROLLABLE CANONICAL FORM
CASE OF A TRANSFER FUNCTION WITH A SIMPLE
NUMERATOR
ecam.fr 47
• Let’s consider the following system
𝒀(𝒔) 𝒃𝒎
𝑯 𝒔 = = 𝒏
𝑼(𝒔) 𝒔 + 𝒂𝟏 𝒔𝒏−𝟏 + ⋯ + 𝒂𝒏−𝟏 𝒔 + 𝒂𝒏
𝑾 𝒔 1
= (B.2)
𝑈(𝑠) 𝑠 𝑛 + 𝑎1 𝑠 𝑛−1 + ⋯ + 𝑎𝑛
𝑌(𝑠)
And is defined as follows
𝑊(𝑠)
𝑌(𝑠) (B.3)
= 𝑏𝑚
𝑊(𝑠)
ecam.fr 48
• From (B.2) we have
𝑠 𝑛 𝑊 𝑠 + 𝑎1 𝑠 𝑛−1 𝑊 𝑠 + ⋯ 𝑎𝑛 𝑊 𝑠 = 𝑈(𝑠)
• Applying the Laplace inverse with initial conditions equal to zeros we get
𝑤 𝑤
𝑤 𝑤
𝑋= 𝑤 𝑋= 𝑤 (B.5)
⋮ ⋮
𝑤 (𝑛−1) 𝑤 (𝑛)
ecam.fr 49
Using (B.4) and (B.5) then we can write
𝒘 ⋯ 𝒘 𝟎
𝟎 𝟏 𝟎
𝒘 𝟏 𝒘 𝟎
⋮ 𝟎 𝟎 𝒘
𝒘 = + 𝟎 𝒖
⋮ ⋮ 𝟎 𝟏 (B.7)
⋮ ⋮ ⋮
−𝒂𝒏 −𝒂𝒏−𝟏 ⋯ −𝒂𝟏
𝒘(𝒏) 𝒘(𝒏−𝟏) 𝟏
Using (B.3) we can write
𝑌 𝑠 = 𝑏𝑚 𝑊(𝑠) (B.8)
• Applying the Laplace inverse to (A.8) we get
𝑦 𝑡 = 𝑏𝑚 𝑤(𝑡)
(B.9)
𝒘
𝒘
𝒚 𝒕 = 𝒃𝒎 𝟎 ⋯ 𝟎 + 𝟎𝒖
⋮
𝒘(𝒏−𝟏)
ecam.fr 50