What I mean by zero minus one, just before we apply the external excitation.

And this tool related to the fact that the power transform is a linear operator.
So you can scale the input, then the power transform is proportional to original
power transform.
This is summation, you combine these two, it's a famous principle of
superposition.
And this frequency shift theory, if your function have t, then you multiply by an
n-weigh of function.
S for minus a t, we call this n-weigh of function.
If a is positive, then the function value together will decrease exponentially.
Then it will sit on the x plane to the left by a of go minus a.
This is important, delay signal.
You signal at t, delay by capital T seconds.
Then the resulting power transform is original power transform of at t, and then
multiply this s for minus s t.
t here is the delay time.
Give you an example, this is a state function, right?
Even if the state function is one, it starts from zero and jumps to one, the t
equal to zero.
The power transform of this state function will be like this.
However, we delay this state function by delta t seconds.
Then the power transform of this delay signal is simply come from original power
transform,
multiply this s for minus delta t is what? Delay time.
Multiply by s.
So this comes from this delay theory.
With this, if you want to obtain the power transform, for example, a pass
function.
Pass function is what?
It's zero, at t equal to zero, and jump to a constant value, and it last for delta
t seconds.
And then come back to zero.
We call this pass function, right?
And the duration time is delta t, the magnitude is one divided by delta t.
Therefore, the area of this pass function is what?
What is the power transform of this pass function?
Of course, you can follow the definition of the power transform.
If the values are integral, you will get a result.
But if you use a concept of delay signal, then this delta pass function t actually
equal to what?
The magnitude is one divided by delta t, step function minus delay step function,
right?
And you can construct this pass function from these two functions, and then you
just change the magnitude.
Okay? So, I tried delay theory, and the power transform of this pass function is
simply equal to what?
The power transform of this one minus the power transform of this one.
What?
So, denominator will be s, and the magnitude here is one minus this value, right?
This is just the magnitude of this function.
And something related to what we wanted to discuss in this lecture is this
differentiation theory.
If you know that the power transform of Ft, then what will be the power transform
of dF dt?
You can always go back to the definition.
But the fact that you don't need to go back to the definition, because
differentiation theory will tell you that
the resulting power transform of dF dt will be s multiplied by Fs.
What is Fs? Fs is the power transform of Ft, right?
 You already know that. Minus initial condition. Zero minus means initial
condition.
Second derivative will be s squared multiplied by Fs,
and minus s multiplied by initial value of the function, and then minus the
changing rate,
the initial value of the changing rate. So Ft is zero minus.
From this, you can see, after you apply the power transform to the differential
equation,
differentiation is replaced by s. So s operator will represent what?
Will represent differentiation.
In other words, one divided by s is an integrator.
So when we talk about the power transform, the dynamic system is very important to
know that s is a differentiation.
One divided by s is called an integrator.
So in your system later, when we talk about transform function, you pay attention
to your denominator.
If you have s in your denominator, that means your system has one integrator.
If your transform function denominator has s squared, that means your system has
two integrators.
You can extend to nth differentiation. These are extensive.
So what does this tell you? If you apply the power transform to dnf, dt power n,
the differentiation operator will be what? s power n multiplied by Fs.
The other thing here, this is polynomial. Polynomial of s.
But determined by what? By initial conditions.
F zero minus, F dot zero minus, and F second derivative zero minus.
But it's just a polynomial.
And this integration theory, if you want to apply the power transform to an
integral of the function Ft,
then the result will be...
Okay.
Okay.
Okay.
We are talking about linear system theory here. We cannot make this work.
So this is small.
So you can see one thing that I have is an integrator.
And this we talk about initial variable theory.
So let's talk about transform function.
We talk about transform function, we also want to talk about response function.
Remember we talk about input-output differential equation, right?
In the system, we identify input variable, output variable,
and we come up with differential equation, and we have this input-output equation
here.
That hand side here includes all the variable, output variable y,
and the first derivative of output variable y dot, up to nth derivative of output
variable.
We put them all on that hand side.
Right hand side here involves input, u dot, and up to nth derivative of input.
So we separate output variable and input variable.
And the hands side here in a way represent the internal dynamics of the system.
And right hand side here represent how the input u comes into the x i o system.
It's not just u coming, the change in rate of u dot also coming.
So remember I explained to you, actually I'm not a fan of this differential
equation
because it's not easy to obtain this differential equation.
And also we call some technique to actually understand the dynamics of this
differential equation.
When we apply the process form, then it becomes much easier.
When we apply the process form to y, we have ys, right?
How about apply to y dot?
 From previous case, we talked about what? Differentiation theory.
So we have this s coming, and we have initial condition coming to this polynomial
here.
So what happens is, apply the process form, we get a polynomial, internal s.
It's s power n, a1 s power n minus 1, according to the coefficient of differential
equation.
This is the internal dynamics.
After applying the process form, we get polynomial internal s.
This represents the internal dynamics.
All the initial conditions come here.
I just represent the initial condition by i s.
You know why I choose i, right?
Initial condition, it's polynomial.
And right hand side here is also polynomial.
So this relationship now is so-called s-grade relationship.
S-grade relationship is input and output.
So this is a result of applying the process form to your new product of
differential equation.
And then we represent, we use ds to represent this polynomial.
ns represents this polynomial.
i s is a polynomial coming from initial condition.
Then we can write very simple in this way, right?
So this is a relationship between input and output and initial condition.
Then we can define so-called response function.
Now you know how to do it because it's just s-grade relationship, right?
You move i s, the right hand side, you change the sign, you divide equation by
denominator polynomial.
Then you will get this output response function.
Approach transform output and two components.
The first component minus one divided by ds,
multiplied by i s, initial condition.
So this is due to initial condition.
Second component is what?
It's ns divided by ds, multiplied by input.
So what does ns mean?
ns just represent how the input coming to excite your internal dynamic.
ds represent the internal dynamics.
And this is called response function.
You apply inverse number of response of this ys, you get the response.
That's the solution, right?
So when you learn differential equation for any time invariant system,
you learn at least two methods.
One is what?
You use a homogeneous equation and get the...
For forcing function, you get particular solution, right?
And then for homogeneous equation, you get complementary solution.
You combine them together, you have complete solution, right?
And you use initial condition to obtain the...
To determine the coefficient, right?
And second method is what?
Use the power transform, like this one.
And then you apply inverse power transform to get the solution.
So the concept of transform function is what?
We ignore initial condition.
Or we assume initial condition ROG.
Then we call this is the transform function of the system.
And in your textbook, usually we put a dynamic system with transform...
Concept of transform function.
We say this is our input inverse.
This is our output bias.
And then we have our transform function TS.
And this is actually because we assume the initial condition.
We ignore the initial condition.
And again, if you look at this transform function here,
remember denominator capitalizer internal dynamic of the system.
New one with capitalizer how the input is what?
Us coming to excite your internal dynamics.
If you are already familiar with transform function,
I believe many of you familiar with transform function.
The important thing associate with denominator is the whole location.
The rules of this polynomial.
And we know whole location is determined by.
Determine the stability of the system.
Determine the dynamic characteristic.
For example, if we have real poles,
then we know the dynamic characteristic of real pole
associated with real pole is exponential function.
A pair of complex country way pole.
If we have pair of complex country way pole,
the dynamic characteristic associated with
pair of complex country way pole is what?
Signs or the function.
Depending on real part, anyway the function.
So that we can then the dynamic characteristic.
But the rules of NS determine what?
For example, rules of the NS new more rate polynomial,
we call it zero.
What does zero mean?
We don't want to talk about composite system.
I will explain to you the examples.
Simply said, it's just how the input coming to excite dynamic,
internal dynamics.
But later I will show you some example.
And hence you understand it about zero of the transform function of the system.
So this is for all related to your system.
I want to give you a number of examples
from the point of view of transform function.
And these examples can be useful to you.
The first one, we also call this single degree of freedom system
with displacement input.
We have lumped mass, spring constant, and dash path.
But our input now is input displacement input yt.
This is a system that requires high speed or motorcycle.
Or you design a vibration isolator.
So the input now is yt.
How to model this simple system?
Again, we use force balance.
We focus on m.
And then you have force equivalent or force balance.
Inertial force plus damping force plus spring force equals zero.
Can everyone see this?
This is just force balance.
But now when you evaluate the damping force,
actual damping force will be what?
The changing way of the displacement,
the difference between this point and this point, right?
The displacement of this point is x.
Displacement of this point is y.
Remember now y is input.
So you have cx dot minus y dot.
Likewise for spring force, it's kx minus yy.
Remember y is an input.
So input and input is no more same.
The important thing is what?
Keep the output variable on the left-hand side.
Input variable on the right-hand side.
Then you do this turn and this turn the other way around.
This is our input output differential equation.
Second part.
You apply the power transform.
Ignore the initial condition.
You get this polynomial.
This will be a denominated polynomial.
This will be a new model polynomial.
Okay?
Then your transform function is like this.
It's not typical spring mesh test process.
Because you have so-called new model dynamics.
Why?
Because your input now is what?
Displacement equal.
Displacement equal.
You call this a second-order system
with new model dynamics.
If you want to design a vibration isolator,
or you want to design a vibration isolator for earthquake,
it's very important.
I mean, in Taiwan, every day we almost have earthquake.
If you look at this one,
you can see C and S will be your design model.
If you want to design C and K,
I should say C and K are your design model.
Because you change C and K,
you change your frequency range.
G of S what?
You will reduce that vibration of your system.
So this becomes design model.
If you design your suspension system
for your motorcycle or your vehicle,
you have four wheelbikes, you have four suspensions.
You want to design your K and C for suspension
to evade your performance.
So this becomes your design model.
Remember this one?
Magnetic suspension system.
And last week we devised an equation
and performed linearization.
After we linearize it,
we get this linear relationship.
I want to remind you what we mean by linearizing.
We need to select a corrective point or equilibrium point.
And that equilibrium point is
I want to have the object at equivalent at XOR,
then I need to have IOR equal to XOR mod.
This comes to me.
We talked about this last week, right?
And with that, then if you add actual small current delta I,
remember now I is IOR plus delta I.
Actual current delta I,
then the response of your displacement
away from the equilibrium point or operating point
is determined by this linear differential equation.
And this differential equation is good
when delta I and delta X are small.
So we get this differential equation.
You can apply the power transform, right?
When you apply the power transform,
you should move this term to the right side.
Right?
All the variable, upper variable,
changing the upper variable has to be upper variable.
And then you apply the power transform,
then you can get this transfer function.
How to get?
Apply the power transform,
I encourage you to do it at home
so you know how to divide to this.
Look at this one.
So this is denominator polynomial, right?
And this new model is again a constant
at the minus sign.
I explain to you why we have minus sign.
Look at this denominator polynomial.
How many roots do we have for this system?
Of course we have two roots, right?
And what is the location of these roots?
And then we have 10 roots for this symbol
called the value of this plus minus sign.
So this symbol is the square root of 10 to the power of 9.
XOR of 9 minus.
So this system has two ports.
One port is stable port.
Another port is unstable port.
In a way, when we say the system is unstable,
for sure this system is unstable.
But actually, this dynamic system has two more.
One more actually stable.
But second, this is more stable.
This one is unstable.
Later when we talk about state space,
we will talk about model decomposition.
It becomes more clear.
But even now you can see the system is not entirely,
I mean it's unstable,
but you have two variables, internal variables,
it's not entirely unstable.
So this is the transfer function.
Okay, if we go back to this system here,
the port will be the mass plant port,
will be the port of 10G.
So this is the meal.
Assuming we have undertaken which of the pair
of companies have paid off, this is the mass plant.
But then we have one zero, right?
We have one zero for this system here.
This zero, depending on your design,
in case it could be closer to origin,
it could be far away from origin.
That could up to your define to determine the zero.
So this is the mass of the zero.
Now let's look at this full mass system.
You have seen this.
You call it constraint structure.
Constraint in earth is like a machine.
You install your machine in the factory
and install the floor of the factory.
This is constraint structure.
We use a simple tool mass, M1 and M2.
And we have one input, Ft.
Then to derive the transfer function,
we start with differential equation,
configuration form.
And we, last week we talked about,
for this example here, we put two equations
into matrix format, right?
And then we apply the transfer function.
We get this edge matrix, edge grid relationship
between input Fs and then two output Y1s and Y2s.
And we have two by two matrix here.
And then we want to evaluate output Y.
So you need to perform inverse matrix.
Inverse matrix of a two by two matrix.
Everyone know how to do it, right?
You need to obtain the determinant.
And then you need to obtain edge matrix.
And then multiply input.
And then you have Ys divided by Fs.
You get two transfer functions.
G1 is for Y1, G2 for Y2.
And delta S is a determinant of this matrix.
It becomes what?
Denominator polynomial.
This denominator polynomial is fourth order polynomial
in terms of S2.
It looks complicated, but it's just fourth order
because we have two mass.
And how to obtain this?
I encourage you to pull it yourself at home.
Okay?
So if we look at the pole location here,
assuming we have under 10, 10 is small.
This is the real part.
This is the edge matrix.
And this was all the polynomial.
Assuming we have two pair of countable conjugate poles,
then it should have a pair of pole values.
And the other pair of pole values is something like this.
Everyone know what that means by countable conjugate.
Explain.
So this is becoming the dynamic characteristic of the system.
You have two pair of countable conjugate poles.
You have double spring mass system,
we call it your two mass.
Look at the linear model here.
This is for G1, this is for G2.
G1 is for associated output of Y1,
and G2 is associated output of G2.
Then it's the same denominator.
It's a quite different new model.
And then if you plot the G2 location for Y2,
you would have a pair of countable conjugate poles.
But if you plot the G4 for G1 associated with Y1,
then it would look something like this.
You have G2 here.
Very different, right?
What does zero mean?
Zero means how the input coming to X1 is dynamic.
In total, dynamic is coming by this pair of countable conjugate poles,
and second pair of countable conjugate poles.
But now, for G2, you have a pair of zero here.
This one, you excite the system,
then this pair of zero almost cancel this pair of poles.
Actually, the concept of canceling,
zero and pole canceling is incorrect.
Usually, we want to think this pair of zero kind of block
the input coming to excite this pair of poles.
That pair of poles over there, you cannot cancel the pole of the system.
Zero just means one.
It's like something tries to stop the input coming to excite this pair of poles.
But for Y1, you don't have a pair of zero very close to this pair of poles.
Instead, you have zero here.
So the dynamic response of N1 and N2 will be quite different.
The concept here, a very important concept, is you apply force to N2, right?
We call this a pole of gradient.
What do you mean by pole of gradient?
You have F applied to N2 directly at N2.
In N1, the actual N1,
but in between you have a spring and you have a damper, right?
We call this long pole of gradient.
Here, it's the same point, but it's the same rigid body.
You apply the force to this one.
If you are working on control or you are working on structural dynamic,
you understand that it's very difficult to control a long pole of gradient.
For example, you go home, try to find a very fragile stick,
and then you hold the stick at one end,
and then you want to control the end point at the other end.
That's so-called long pole of gradient control.
It's very difficult to control.
Even if you use your eye to do feedback, try to control it very difficult,
because of the flexibility in the grid.
About 25 years ago, this space scope, the Hubble scope,
actually when they sent it to outer space,
they told them they cannot get very clear picture of the space.
The reason is because the lens is not light, perfect.
So they have astronauts try to bring a very lightweight lobe,
and try to use the lobe to just move the lens a little bit.
That lobe is just a stick, a very long stick.
The astronaut is at one end, and the lens is at the other end.
It took almost 10 hours to try to fix that small thing.
It's because we are doing long pole of gradient control.
Very difficult to do that.
So here, for us, I just want to mention to you,
you can say I get a mathematical, and this is a force auto dynamics,
but I just want to point out to you,
depending on which output you're talking about,
the dynamic can be very difficult.
Now let's look at this time constraint structure.
The constraint structure of the machine on Earth,
unconstraint structure that the airplane, like the space vehicle,
it's not constraint.
We also used two mass, m1 and m2,
but now we have two force, f1 and f2.
So now we have two input and two output systems.
And then we write down the equation, put it in the management form,
it looks very similar, right?
But it just result constraint.
And then this is the applied approach,
and again we get a two by two matrix.
But now we have f1 and f2.
How to get the transfer function?
You will get transfer function matrix.
You will get two inputs, two output.
You inverse matrix,
and then you can find so-called transfer function matrix.
We have ys divided by fs.
This is very strange.
This is a vector, right?
But which commanding is two by two matrix means what?
g1,1 means what?
Input is what?
Output is what?
Input f1, how the f1 effect output y1?
g2,1 is what?
How the input f1 effect output y2?
And g1,2 is what?
How the input f2 effects output y1?
And this is f2 effect y2.
So in other words, this vector y is equal to this transfer function matrix,
multiplied by input vector.
It will be something like this.
So input output is actually a compound.
This output vector is equal to this two by two matrix.
Transfer function matrix, multiplied by f1,s.
So the physical meaning of g1,1 is what?
g1,1 multiplied by f1, in to y1.
So you want to understand the physical meaning of g1,1, g2,1, g2,2.
So the denominator polynomial is a determinant of this two by two matrix.
When I say determinant, everyone understands what I'm saying, right?
Just this guy right here, this thing, and then subtract this thing.
Become force order polynomial.
And then this is a numerator.
Look at this polynomial.
What do you guys got?
Force order polynomial.
But this force order polynomial can be factorized like this.
What do you use this for?
You have m1, m2, m3.
The mod of the second order polynomial.
The second order polynomial to do not is not a stress.
It should be very clear.
The mean of the second order polynomial is a simple square matrix.
But what does s square mean?
This is the denominator.
Earlier, we made a problem.
What do we mean by denominator?
The integrator, right?
So now we have a double integrator.
Why we have a double integrator?
Because it's an unconstrained system.
An unconstrained system like that.
If your space vehicle is an interior space vehicle in a simple sense,
no constraints.
I apply a force.
What is the difference between the equation on this one?
So we use x and let's say y.
y double dot is equal to n times n pi.
This is a differential equation for unconstrained system like a space vehicle.
What is a transfer function?
This is a transfer function.
And then...
This is a space vehicle.
It's like a dumb mess.
You apply force, then you have constant acceleration.
So your output will be low.
You integrate acceleration once, you get velocity.
You integrate two times, you have position y, right?
So it's a double integrator.
For our unconstrained system here,
if you look at the denominator,
remember this is our ds, okay?
You've got to have s equal to this unconstrained.
Unconstructed means what?
M1 and M2, they will move together.
Like a bridge body, right?
How about the secondary of this?
It's like an elastic system with a spring,
with mass, spring and that's what.
It represents what?
The relative amplitude of the number.
So, why I have these two examples,
I want to show you, you want to be very careful
when you develop your mathematical model.
Without constraint, and with constraint,
without constraint, you definitely need to get s squared.
If you don't get s squared, you cannot factor out s squared.
The only reason for it is by mistake.
So that depends on constraint.
You can't get it wrong.
So it's not here you're just trying to derive
the differential equation and then obtain the transfer function.
You want to look at your transfer function,
you want to look at your denominator,
you want to be able to explain why you get this result.
Because you anticipate this kind of result.
For the circuit, this downward circuit,
I can use the charge in C1 and charge in C2
with the two variables.
And then we can write down the differential equation.
And then the product voltage one.
And then this is our zero equal to dy.
So we have our transfer function.
We have the transfer function.
 G1 associated with Q1.
G2 associated with Q2.
When you see something similar,
look at the numerator,
the second order denominator here for Q1.
And then for Q2, you have zero as a function.
This is something that gets collocated and uncollocated.
Very similar to the mechanical system.
Collocated and uncollocated.
So in other words, these are voltage source
trying to charge to this capacitor, right?
And this is like collocated excitation.
This one can be charged, but it's uncollocated.
Some of you probably don't care about collocated, uncollocated,
but from the point of view of dynamics, they can be quite different.
They don't want to talk about frequency response.
They may be wrong.
Okay, so these are transfer function differential equations.
So you all need to start from configuration.
You need to write down the number of differential equations.
And if you have a couple of them,
you need to put in a matrix form,
and then you write a construct form,
then you manipulate this result in a matrix as wave relationship,
then you can get transfer function matrix.
Now we want to talk about state-space communication.
Actually, this will be a topic in chapter 4, chapter 4, 5, 6, 7.
We will focus on state-space communication.
Here I just want to talk about how many different wave present as a matter of
matter, right?
So this is one of the wave, okay?
So the difference now, so far we talked about input, output,
we talked about transfer function.
We focus on input variable and output variable.
But in addition to input variable, output variable,
the system is still, we have so-called internal variable.
So we want to have input variable, output variable,
and then we want to have state variable in between.
Actually, this state variable completely determines the state of the dynamic
system.
Remember when we talk about input variable,
for real system, input variable comes from actuator, right?
How many actuators come in your system?
When we say we are dumb, we are mad system, we are F1, F2.
F1, F2 has to come from actuator, right?
You have that rocket, then you never have actuator, right?
Output variable actually comes from measure.
It's a sense of measure, okay?
But internal variable actually determines the dynamic of the system.
The concept of state-space communication is we need to determine a set of state
variables.
And this set of state variables is the smallest set of independent variables.
They have to be independent, okay?
And they completely describes the entire state of the dynamic system.
You might say, what do you mean by state of dynamic system?
Means if you know the state of dynamic at this time, Tg log,
and then you apply expectations from Tg log,
then you can know, you can predict whatever of the dynamic system come into play,
okay?
 So this is what state variable is.
So this is a small set of independent variable completely described the entire
state.
If you go back to input-output relationship,
say you have nth order differential course,
how many initial conditions you need to determine your solution?
You need to have n initial conditions.
You have y, g of minus, y plus, g of minus, up to y.
And minus one, then you get the initial time, zero minus one.
You get the n of y, y.
So from the framework of context,
you need to add state variable to completely determine your dynamic state.
So this is a mass, minimum number of state variables.
One thing about state-space generation is we have input variable, output variable,
and we have state variable.
You can put them in a measure form.
We call that vector generation.
We use a actually, we call it state vector.
This is a vector.
Actually consists of all the state variables.
This one is a vector.
So this is the m i one.
This is the column vector.
And then we have a u t is our input vector.
So this is a x, this is u.
This is the p of it.
We call that input vector.
And then our output vector is y vector.
Depending on the output, we have y for minimum.
We have q output.
So we assume we have p input, we have q output,
but we have n state variable.
We put them into a vector form.
Then what will be our mass?
We want to display the dynamic of the system.
We want to know how the state variable is changing the state.
So we want to know x star.
So in state-space generation, instead of using output variable,
we want to divide the relationship of dynamic up to n state variables.
In state-value generation, we just need to know x1 dot, x2 dot, x3 dot.
These are changing the state variable.
And in vector form, we can have a very simple interpretation.
The fact is that if you represent a state variable as a vector,
the dynamics of the system will be represented by these state vectors.
And this state vector will be a function of time for variable t.
And this time for variable t, if you imagine a movement in n-dimensional vector
space,
it will be a vector in a vector space, and this vector changes its time.
And the vector is like a trajectory.
It changes like a trajectory in your vector space.
Assuming you have a three-state variable with three-dimensional vector space,
then your dynamic is like a motion in the three-dimensional space.
It moves with time.
So it forms a state space. It's a vector space.
Later on, in chapter four, we will talk about vector space.
Here I just want to mention the actual dynamics that move in the n-dimensional
vector space.
So the model now can be represented as this.
This is our measurement model for the dynamic system.
Dynamic system is what we're changing with.
You want to know at this moment, at this state of the dynamic system,
what will be the changing way of the state of the dynamic?
And then H1 dot, it should be function of t.
H1 dot will be function of all the state variables.
You call it a couple of variables.
All the input variables, you make sense, right?
Input coming, state variable itself will determine your changing way,
but input will also determine your changing way.
It could also depend on time t.
We have time varying system.
This is just a general representation.
This is a concept.
Say our measurement model try to determine this function H1 for H1 dot.
We want to determine H2 for H2 dot up to Hn for Hn dot.
So state variable representation, nothing more than trying to determine
the changing way of each individual state variable.
But they are coupled with each other.
That's why we say H1 is function of n state variable, p input variable,
and n pole variable.
And this function H1 could be nonlinear.
And you can even, I mean, you start with very simple spring-mass
state-space system.
It's linear function, right?
But in general, it could be nonlinear.
So this is our measurement model.
We call this state equations.
And then our output variable, q output y1, output y1 could be a function
of all the state variables, input variables.
The difference between output equation, computer state equation,
it's a major difference.
Can someone tell me what's the difference between state equation
and output equation?
You see that linear function down.
What kind of equation we have here?
It's differential equation, right?
It just has n of them coupled together.
How about the output equation?
What kind of equation we have?
We call this air-spray equation.
It's just a function, right?
It won't change anything.
Air-spray equation.
I really want to encourage you guys to review the coming hours
so when you discuss this, everyone uses the same terminology.
It's easier to communicate.
You guys learn a lot of techniques,
but you don't focus on terminology.
And when I say, like I want to mention to you,
this is a differential equation, this kind of equation.
It's very important, right?
Because what?
Output is just output measurement.
Measurement of the result.
The result has to be a function of your internal variable, right?
So it's an air-spray equation.
But state equation is changing with time
because we are talking about dynamic system, right?
So we want to know the changing weight.
 So it's a differential equation.
If we put in the vector of H, then it will go down like this.
The x star equals H. X, U, T.
H is what?
The vector function.
Okay.
How many components of H?
It's a vector of what?
N of what?
Y is an air-spray function of U, the X, U, T.
G is a vector.
How many of them?
Q of them, right?
So U is P by 1.
X is M by 1.
Y is Q by 1.
Because we assume we have P input, Q output, and N state variables.
Okay?
P input, N state variable, Q output.
And this is talking about...
You might pay attention.
I already...
Just to give correct model, you need a unique solution.
This equation here is the unique solution.
Correct way to generate dynamic system.
So this is a general space case.
In general, of course, we assume we are changing weight
to the domino function of state variable U, the vector of what?
We don't have to deal with the domino system.
We hope in our dynamic system, we do.
Right?
So what do you mean by domino?
Earlier we said of changing weight, X star, domino function of X, domino function
of U, what?
You want to say we have a linear system.
That's only when this relationship is linear.
Which means what?
X star will be the linear function of input variable and state variable.
So for linear time variance system is
X star P changing weight of your state variable
over time of weighting
will be linear function, linear relation A mod X, E mod U.
How to vary the relationship?
I think I should stop here.
I'm sitting here all the time already.
I need to finish this chapter.
I'm going to finish this one.
Someone should stop me for some time.
Thank you.