student Workbook

Inverted Pendulum Experiment for LabVIEW ™ Users

Standardized for ABET * Evaluation Criteria
Developed by:
Jacob Apkarian, Ph.D., Quanser
Paul Karam, B.A.SC., Quanser
Michel Lévis, M.A.SC., Quanser

Quanser educational solutions

are powered by:

Course material
complies with:

Captivate. Motivate. Graduate.

*ABET Inc., is the recognized accreditor for college and university programs in applied science, computing, engineering, and technology. ABET has provided leadership
and quality assurance in higher education for over 75 years.
© 2012 Quanser Inc., All rights reserved.

Quanser Inc.
119 Spy Court
Markham, Ontario
L3R 5H6
Canada
[email protected]
Phone: 1-905-940-3575
Fax: 1-905-940-3576

Printed in Markham, Ontario.

For more information on the solutions Quanser Inc. offers, please visit the web site at:
http://www.quanser.com

This document and the software described in it are provided subject to a license agreement. Neither the software nor this document may be
used or copied except as specified under the terms of that license agreement. All rights are reserved and no part may be reproduced, stored in
a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior
written permission of Quanser Inc.

ACKNOWLEDGEMENTS
Quanser, Inc. would like to thank the following contributors:

Dr. Hakan Gurocak, Washington State University Vancouver, USA, for his help to include embedded outcomes assessment,

Dr. K. J. Åström, Lund University, Lund, Sweden for his contributions to energy-based control, and

Andy Chang, National Instruments, Austin, Texas, for his help in designing this lab.

ROTPEN Workbook - Student Version 2

CONTENTS
1 Introduction 4

2 Modeling 5
2.1 Background 5
2.2 Pre-Lab Questions 8
2.3 In-Lab Exercises 9
2.4 Results 13

3 Balance Control 14
3.1 Specifications 14
3.2 Background 14
3.3 Pre-Lab Questions 18
3.4 In-Lab Exercises 19
3.5 Results 24

4 Swing-Up Control 25
4.1 Background 25
4.2 Pre-lab Questions 27
4.3 In-lab Exercises 28
4.4 Results 30

5 System Requirements 31
5.1 Overview of Files 31
5.2 Hardware Setup 32
5.3 Software Setup 32

6 Lab Report 34
6.1 Template for Content (Modeling) 34
6.2 Template for Content (Balance Control Experiment) 35
6.3 Template for Content (Swing-Up Control Experiment) 36
6.4 Tips for Report Format 37

ROTPEN Workbook - Student Version v 1.1

1 INTRODUCTION
The objective of this laboratory is to design and implement a state-feedback control system that will balance the
pendulum in the upright, vertical position.

Topics Covered

• Linearizing nonlinear equations of motion.

• Obtaining the linear state-space representation of the rotary pendulum plant.

• Designing a state-feedback control system that balances the pendulum in its upright vertical position using
Pole Placement.
• Simulating the closed-loop system to ensure the specifications are met.

• Introduction to a nonlinear, energy-based swing up control.

• Implementing the controllers on the Quanser SRV02 Rotary Pendulum plant and evaluating its performance.

Prerequisites

• Know the basics of LabVIEW™ .

• Understand state-space modeling fundamentals.

• Some knowledge of state-feedback.

ROTPEN Workbook - Student Version 4

2 MODELING
2.1 Background

2.1.1 Model Convention

The rotary inverted pendulum model is shown in Figure 2.1. The rotary arm pivot is attached to the SRV02 system
and is actuated. The arm has a length of Lr , a moment of inertia of Jr , and its angle, θ, increases positively when it
rotates counter-clockwise (CCW). The servo (and thus the arm) should turn in the CCW direction when the control
voltage is positive, i.e., Vm > 0.
L
The pendulum link is connected to the end of the rotary arm. It has a total length of Lp and it center of mass is 2p .
The moment of inertia about its center of mass is Jp . The inverted pendulum angle, α, is zero when it is perfectly
upright in the vertical position and increases positively when rotated CCW.

Figure 2.1: Rotary inverted pendulum conventions

2.1.2 Nonlinear Equations of Motion

Instead of using classical mechanics, the Lagrange method is used to find the equations of motion of the system.
This systematic method is often used for more complicated systems such as robot manipulators with multiple joints.

More specifically, the equations that describe the motions of the rotary arm and the pendulum with respect to the
servo motor voltage, i.e. the dynamics, will be obtained using the Euler-Lagrange equation:

∂2L ∂L
− = Qi
∂t∂ q˙i ∂qi

ROTPEN Workbook - Student Version v 1.1

The variables qi are called generalized coordinates. For this system let

q(t)⊤ = [θ(t) α(t)] (2.1)

where, as shown in Figure 2.1, θ(t) is the rotary arm angle and α(t) is the inverted pendulum angle. The corre-
sponding velocities are [ ]
⊤ ∂θ(t) ∂α(t)
q̇(t) =
∂t ∂t

dθ
Note: The dot convention for the time derivative will be used throughout this document, i.e., θ̇ = dt . The time
variable t will also be dropped from θ and α, i.e., θ = θ(t) and α = α(t).

With the generalized coordinates defined, the Euler-Lagrange equations for the rotary pendulum system are

∂2L ∂L
− = Q1
∂t∂ θ̇ ∂θ
∂2L ∂L
− = Q2
∂t∂ α̇ ∂α

The Lagrangian of the system is described

L=T −V
where T is the total kinetic energy of the system and V is the total potential energy of the system. Thus the Lagrangian
is the difference between a system's kinetic and potential energies.

The generalized forces Qi are used to describe the non-conservative forces (e.g., friction) applied to a system with
respect to the generalized coordinates. In this case, the generalized force acting on the rotary arm is

Q1 = τ − Br θ̇

and acting on the pendulum is

Q2 = −Bp α̇.

See [2] for a description of the corresponding SRV02 parameters (e.g. such as the back-emf constant, km ). Our
control variable is the input servo motor voltage, Vm . Opposing the applied torque is the viscous friction torque, or
viscous damping, corresponding to the term Br . Since the pendulum is not actuated, the only force acting on the
link is the damping. The viscous damping coefficient of the pendulum is denoted by Bp .

The Euler-Lagrange equations is a systematic method of finding the equations of motion, i.e., EOMs, of a system.
Once the kinetic and potential energy are obtained and the Lagrangian is found, then the task is to compute various
derivatives to get the EOMs. After going through this process, the nonlinear equations of motion for the SRV02
rotary inverted pendulum are:
( ) ( )
1 1 1
mp Lr + mp Lp − mp Lp cos(α) + Jr θ̈ −
2 2 2 2
mp Lp Lr cos(α) α̈
4 4 2
( ) ( )
1 1
+ 2
mp Lp sin(α) cos(α) θ̇α̇ + mp Lp Lr sin(α) α̇2 = τ − Br θ̇ (2.2)
2 2
( )
1 1 1
− mp Lp Lr cos(α)θ̈ + Jp + mp L2p α̈ − mp L2p cos(α) sin(α)θ̇2
2 4 4
1
− mp Lp g sin(α) = −Bp α̇. (2.3)
2

The torque applied at the base of the rotary arm (i.e., at the load gear) is generated by the servo motor as described
by the equation
ηg Kg ηm kt (Vm − Kg km θ̇)
τ= . (2.4)
Rm

ROTPEN Workbook - Student Version 6

See [2] for a description of the corresponding SRV02 parameters (e.g. such as the back-emf constant, km ).

Both the equations match the typical form of an EOM for a single body:

J ẍ + bẋ + g(x) = τ1

where x is an angular position, J is the moment of inertia, b is the damping, g(x) is the gravitational function, and τ1
is the applied torque (scalar value).

For a generalized coordinate vector q, this can be generalized into the matrix form

D(q)q̈ + C(q, q̇)q̇ + g(q) = τ (2.5)

where D is the inertial matrix, C is the damping matrix, g(q) is the gravitational vector, and τ is the applied torque
vector.

The nonlinear equations of motion given in 2.2 and 2.3 can be placed into this matrix format.

2.1.3 Linearizing

Here is an example of how to linearize a two-variable nonlinear function called f (z). Variable z is defined

z ⊤ = [z1 z2 ]

and f (z) is to be linearized about the operating point

z0 ⊤ = [a b]

The linearized function is

( ) ( )
∂f (z) ∂f (z)
flin = f (z0 ) + (z1 − a) + (z2 − b)
∂z1 z=z0 ∂z2 z=z0

2.1.4 Linear State-Space Model

The linear state-space equations are

ẋ = Ax + Bu (2.6)
and
y = Cx + Du (2.7)
where x is the state, u is the control input, A, B, C, and D are state-space matrices. For the rotary pendulum system,
the state and output are defined
x⊤ = [θ α θ̇ α̇] (2.8)
and
y ⊤ = [x1 x2 ]. (2.9)

In the output equation, only the position of the servo and link angles are being measured. Based on this, the C and
D matrices in the output equation are [ ]
1 0 0 0
C= (2.10)
0 1 0 0
and [ ]
0
D= . (2.11)
0

The velocities of the servo and pendulum angles can be computed in the digital controller, e.g., by taking the deriva-
tive and filtering the result though a high-pass filter.

ROTPEN Workbook - Student Version v 1.1

2.2 Pre-Lab Questions
1. Linearize the first nonlinear inverted rotary pendulum equation, Equation 2.2. The initial conditions for all the
variables are zero, i.e., θ0 = 0, α0 = 0, θ̇0 = 0, α̇0 = 0.

2. Linearize the second nonlinear inverted rotary pendulum equation, Equation 2.3, with initial conditions θ0 = 0,
α0 = 0, θ̇0 = 0, α̇0 = 0.

3. Fit the two linear equations of motion found in the above exercises into the matrix form shown in Equation 2.5.
Make sure the equation is in terms of θ and α (and its derivatives).

4. Solve for the acceleration terms in the equations of motion. You can either solve this using the two linear
equations or using the matrix form. If you're doing it in the matrix form, recall that the inverse of a 2x2 matrix is
[ ]−1 [ ]
a b 1 d −b
A−1 = = , (2.12)
c d det(A) −c a

where det(A) = ad − bc.

In any case, you'll have two equations of the form: θ̈ = g1 (θ, α, θ̇, α̇) and α̈ = g2 (θ, α, θ̇, α̇). Make sure you
collect the terms with respect to the θ, α, θ̇, and α̇ variables.

5. Find the linear state-space of the rotary inverted pendulum system. Make sure you give the A and B matrices
(C and D have already been given in Section 2.1).

ROTPEN Workbook - Student Version 8

2.3 In-Lab Exercises
Before going through these exercises, make sure to open the LabVIEW project called Quanser SRV02 Rotary
Pendulum (Student), pictured in Figure 2.2.

Figure 2.2: Rotary Pendulum LabVIEW project

2.3.1 Simulation: Model Analysis

1. Under the Control Design and Simulation folder, open ROTPEN Modeling (Student) VI shown in Figure 2.3.
2. Go the VI block diagram. As shown in Figure 2.4, it is not complete. The MathScript node contains the following
code:

% State Space Representation

A = eye(4,4);
B = [0;0;0;1];
C = eye(2,4);
D = zeros(2,1);

% Add actuator dynamics

A(3,3) = A(3,3) - Kg^2*kt*km/Rm*B(3);
A(4,3) = A(4,3) - Kg^2*kt*km/Rm*B(4);
B = Kg * kt * B / Rm;

system = ss(A,B,C,D);

The representative C and D matrices have already been included. You need to enter the state-space matrices
A and B that you found in Section 2.2. Notice the actuator dynamics have been added to convert your state-
space matrices to be in terms of voltage. Recall that the input of the state-space model you found in Section
2.2 is the torque acting at the servo load gear (or the pivot of the pendulum). However, we do not control
torque directly - we control the servo input voltage. The above code uses the voltage-torque relationship given
in Equation 2.4 in Section 2.1.2 to tranform torque to voltage.

ROTPEN Workbook - Student Version v 1.1

 Figure 2.3: Front panel of ROTPEN Modeling.vi

Figure 2.4: Block diagram of ROTPEN Modeling.vi

3. Enter the correct values in the Rotary Model Parameters control and run the VI. You can find these values in
[3]. The VI is shown running with the default parameter values in Figure 2.5. Attach to your report a screen
capture of the model when the VI is running and enter the numeric state-space matrices in Table 2.1.
4. Record the open-loop poles of the system in Table 2.1.

5. Enter a name in Model Name and click on the OK button to save the model. This model will later be used with
other VIs.

Before ending this lab... To do the pre-lab questions in Section 3.3, you need the A and B matrices (numerical
representation) and the open-loop poles. Make sure you record these.

2.3.2 Implementation: Calibration

1. In the Quanser SRV02 Rotary Pendulum LabVIEW project, open the ROTPEN Control (Student) VI. Make
sure the HIL Initialize is configured for your data acquisition device, as explained in Section 5.3.

ROTPEN Workbook - Student Version 10

 Figure 2.5: Running ROTPEN Modeling.vi with default values

2. Turn ON the power amplifier.

3. Run the VI. It should look similarly as shown in Figure 2.6.

4. Rotate the arm and the pendulum in the counter-clockwise direction and examine the direction of their re-
sponse. Does the direction of these measurements agree with the modeling conventions given in Section
2.1.1? Explain why or why not.

5. Click on the STOP button to stop the VI.

6. Go to the VI block diagram. As shown in Figure 2.7, the block diagram is not completed.

7. Connect a control called u (V), which denotes the control input, to the analog output/write VI (remove the
constant 0). When you enter 1 V (as the VI is running), ensure the rotary arm is moving according to the model
conventions that were defined in Section 2.1.1. Attach the response you obtain to your report. Use the Gain
or Multiply VI if changes are necessary.

8. Make sure the VI is stopped by clicking on the STOP button

9. Close the VI when done.

10. Shut off the power amplifier.

ROTPEN Workbook - Student Version v 1.1

 Figure 2.6: Running the ROTPEN Control.vi

Figure 2.7: Incomplete ROTPEN Control.vi block diagram

ROTPEN Workbook - Student Version 12

2.4 Results
Fill out Table 2.1 with your answers from your modeling lab results - both simulation and implementation.

Description Symbol Value Units

State-Space Matrix A

State-Space Matrix B

State-Space Matrix C

State-Space Matrix D

Open-loop poles OL

Table 2.1: Results

ROTPEN Workbook - Student Version v 1.1

3 BALANCE CONTROL
3.1 Specifications
The control design and time-response requirements are:

Specification 1: Damping ratio: ζ = 0.7.

Specification 2: Natural frequency: ωn = 4 rad/s.
Specification 3: Maximum pendulum angle deflection: |α| < 15 deg.
Specification 4: Maximum control effort / voltage: |Vm | < 10 V.

The necessary closed-loop poles are found from specifications 1 and 2. The pendulum deflection and control effort
requirements (i.e., specifications 3 and 4) are to be satisfied when the rotary arm is tracking a ±20 degree angle
square wave.

3.2 Background
In Section 2, we found a linear state-state space model that represents the inverted rotary pendulum system. This
model is used to investigate the inverted pendulum stability properties in Section 3.2.1. In Section 3.2.2, the notion
of controllabitliy is introduced. The procedure to transform matrices to their companion form is described in Section
3.2.3. Once in their companion form, it is easier to design a gain according to the pole-placement principles, which is
discussed in Section 3.2.4. Lastly, Section 3.2.6 describes the state-feedback control used to balance the pendulum.

3.2.1 Stability

The stability of a system can be determined from its poles ([6]):

• Stable systems have poles only in the left-hand plane.

• Unstable systems have at least one pole in the right-hand plane and/or poles of multiplicity greater than 1 on
the imaginary axis.
• Marginally stable systems have one pole on the imaginary axis and the other poles in the left-hand plane.

The poles are the roots of the system's characteristic equation. From the state-space, the characteristic equation of
the system can be found using
det (sI − A) = 0
where det() is the determinant function, s is the Laplace operator, and I the identity matrix. These are the eigenvalues
of the state-space matrix A.

3.2.2 Controllability

If the control input u of a system can take each state variable, xi where i = 1 . . . n, from an initial state to a final state
then the system is controllable, otherwise it is uncontrollable ([6]).

Rank Test The system is controllable if the rank of its controllability matrix
[ ]
T = B AB A2 B . . . An B (3.1)
equals the number of states in the system,
rank(T ) = n.

ROTPEN Workbook - Student Version 14

3.2.3 Companion Matrix

If (A, B) are controllable and B is n × 1, then A is similar to a companion matrix ([1]). Let the characteristic equation
of A be
sn + an sn−1 + . . . + a1 .

Then the companion matrices of A and B are

 
0 1 ··· 0 0
 0 0 ··· 0 0 
 
 .. .. .. .. .. 
à =  . . . . .  (3.2)
 
 0 0 ··· 0 1 
−a1 −a2 · · · −an−1 −an

and  
0
 .. 
 
B̃ =  .  (3.3)
0
1

Define
W = T T̃ −1
where T is the controllability matrix defined in Equation 3.1 and

T̃ = [B̃ B̃ Ã . . . B̃ Ãn ].

Then
W −1 AW = Ã
and
W −1 B = B̃.

3.2.4 Pole Placement

If (A,B) are controllable, then pole placement can be used to design the controller. Given the control law u = −Kx,
the state-space in Equation 2.6 becomes

ẋ = Ax + B(−Kx)
= (A − BK)x

To illustate how to design gain K, consider the following system

 
0 1 0
A = 0 0 1 (3.4)
3 −1 −5

and  
0
B = 0 (3.5)
1

Note that A and B are already in the companion form. We want the closed-loop poles to be at [−1 − 2 − 3]. The
desired characteristic equation is therefore

(s + 1)(s + 2)(s + 3) = s3 + 6s2 + 11s + 6 (3.6)

ROTPEN Workbook - Student Version v 1.1

For the gain K = [k1 k2 k3 ], apply control u = −Kx and get
 
0 1 0
A − KB =  0 0 1 .
3 − k1 −1 − k2 −5 − k3

The characteristic equation of A − KB is

s3 + (k3 + 5)s2 + (k2 + 1)s + (k1 − 3) (3.7)

Equating the coefficients between Equation 3.7 and the desired polynomial in Equation 3.6
k1 − 3 = 6
k2 + 1 = 11
k3 + 5 = 6

Solving for the gains, we find that a gain of K = [9 10 1] is required to move the poles to their desired location.

We can generalize the procedure to design a gain K for a controllable (A,B) system as follows:

Step 1 Find the companion matrices à and B̃. Compute W = T T̃ −1 .

Step 2 Compute K̃ to assign the poles of à − B̃ K̃ to the desired locations. Applying the control law u = −Kx to
the general system given in Equation 3.2,
 
0 1 ··· 0 0
 0 0 ··· 0 0 
 
 . . . . . 
à =  .. .. .. .. ..  (3.8)
 
 0 0 ··· 0 1 
−a1 − k1 −a2 − k2 · · · −an−1 − kn−1 −an − kn

Step 3 Find K = K̃W −1 to get the feedback gain for the original system (A,B).

Remark: It is important to do the K̃ → K conversion. Remember that (A,B) represents the actual system while the
companion matrices à and B̃ do not.

3.2.5 Desired Poles

The rotary inverted pendulum system has four poles. As depicted in Figure 3.1, poles p1 and p2 are the complex
conjugate dominant poles and are chosen to satisfy the natural frequency, ωn , and damping ratio, ζ, specifications
given in Section 3.1. Let the conjugate poles be
p1 = −σ + jωd (3.9)
and
p2 = −σ − jωd (3.10)
√
where σ = ζωn and ωd = ωn 1 − ζ is the damped natural frequency. The remaining closed-loop poles, p3 and p4 ,
2

are placed along the real-axis to the left of the dominant poles, as shown in Figure 3.1.

3.2.6 Feedback Control

The feedback control loop that balances the rotary pendulum is illustrated in Figure 3.2. The reference state is
defined
xd = [θd 0 0 0]

ROTPEN Workbook - Student Version 16

 Figure 3.1: Desired closed-loop pole locations

where θd is the desired rotary arm angle. The controller is

u = K(xd − x). (3.11)

Note that if xd = 0 then u = −Kx, which is the control used in the pole-placement algorithm.

Figure 3.2: State-feedback control loop

When running this on the actual system, the pendulum begins in the hanging, downward position. We only want the
balance control to be enabled when the pendulum is brought up around its upright vertical position. The controller
is therefore {
K(xd − x) |x2 | < ϵ
u=
0 otherwise
where ϵ is the angle about which the controller should engage. For example if ϵ = 10 degrees, then the control will
begin when the pendulum is within ±10 degrees of its upright position, i.e., when |x2 | < 10 degrees.

ROTPEN Workbook - Student Version v 1.1

3.3 Pre-Lab Questions
1. Based on your analysis in Section 2.3, is the system stable, marginally stable, or unstable? Did you expect
the stability of the inverted pendulum to be as what was determined?

2. Using the open-loop poles, find the characteristic equation of A.

3. Give the corresponding companion matrices à and B̃. Do not compute the transformation matrix W (this will
be done in the lab using LabVIEW™ ).
4. Find the location of the two dominant poles, p1 and p2 , based on the specifications given in Section 3.1. Place
the other poles at p3 = −30 and p4 = −40. Finally, give the desired characteristic equation.
5. When applying the control u = −K̃x to the companion form, it changes (Ã, B̃) to (Ã − B̃ K̃, B̃). Find the gain
K̃ that assigns the poles to their new desired location.

ROTPEN Workbook - Student Version 18

3.4 In-Lab Exercises
Before going through these exercises, make sure you open the LabVIEW project called Rotary Pendulum (Student),
pictured in Figure 2.2.

3.4.1 Control Design

1. Under the Control Design and Simulation folder, open ROTPEN Control Design (Student) VI shown in Figure
3.3.

Figure 3.3: Front panel of ROTPEN Control Design.vi

2. Run the VI. It should look similarly as shown in Figure 3.4 (except using your model). The VI will prompt you
to find a model file (unless you selected a model in File Path aleady). Find the model you saved from the
previous modeling lab.

Figure 3.4: ROTPEN Control Design (Student).vi when running

3. Go to the Controllability Test tab. Is the system controllable? Explain why.

4. Click on the Companion tab. The Companion matrices à and B̃ for the model loaded are automatically found
(denoted as Ac and Bc in LabVIEW). In order to find the gain K, we need to find the transformation matrix

ROTPEN Workbook - Student Version v 1.1

 W = T T̃ −1 (note: T̃ is denoted as Tc in the VI). Stop the VI. In the VI block diagram, go to the Companion
case. As shown in Figure 3.5, the block diagram is incomplete. Using the VIs from the Diagram Disabled
Structure, calculate the controllability matrix T , the companion controllabilty matrix T c, and the inverse of T c.
Show your completed block diagram and the resulting T , T c, Inverse Tc, and W matrices.

Figure 3.5: Incomplete block diagram of Companion tab

5. Enter the companion gain, K̃, you found in the pre-lab in the Companion Gain: Kc input box on the VI front
panel. Run the VI again to calculate the feedback gain K and record its value in Table 3.1.
6. Click on the CL Plant Analysis. This shows the closed-loop poles of the system, i.e., the eigenvalues of A−BK.
Record the closed-loop poles of the system when using the gain K calculated above. Have the poles been
placed to their desired locations? If not, then go back and re-investigate your control design until you find a
gain that positions the poles to the required location.
7. Select the Pole Placement Design. No gain is generated because, as shown in Figure 3.6, the block diagram is
not complete. Back in the Companion tab, the necessary matrices to find gain K were found manually through
matrix operations. All that work can instead be done using the Pole Placement VI. Add the Pole Placement VI
to the block diagram so the gain is automatically generated from the state-space model and the desired poles.
The Pole Placement VI is found in the Control Design palette. Run the VI again and verify that the gain is the
same as generated before.

Figure 3.6: Incomplete block diagram of Pole Placement tab

ROTPEN Workbook - Student Version 20

3.4.2 Simulating the Balance Control

1. Under the ROTPEN Control Design and Simulation folder, open ROTPEN Control Simulation (Student) VI
shown in Figure 3.7.

Figure 3.7: Front panel of ROTPEN Control Simulation (Student).vi

2. Go to the block diagram. As depicted in Figure 3.8, the block diagram is not finished. Add the necessary
Gain and State-Space VIs to complete the feedback loop (remove the dummy state and control input constant
arrays). In order to use your inverted pendulum model, make sure you use the model that is loaded from Read
Model from File VI.
Hint: Because you are using a vector gain, make sure you set the Polymorphic instance of the Gain VI to
Vector,MatrixGain.

Figure 3.8: Incomplete block diagram of ROTPEN Control Simulation (Student).vi

3. Go to the front panel of the VI and enter the gain K you found in Section 3.4.1.
4. Run the VI. The VI will prompt you to find a model file (unless you selected a model in File Path aleady). Find
the model you saved in the Modeling Laboratory in Section 2.3. The VI is shown running with an arbitrary gain
in Figure 3.9. Include the response of the VI running with your designed gain and attach it to your report.

ROTPEN Workbook - Student Version v 1.1

 Figure 3.9: ROTPEN Control Simulation.vi when running

5. Measure the pendulum deflection and voltage used. Are the specifications given in Section 3.1 satisfied?
6. Press on the STOP button and close the VI when you are done.

3.4.3 Implementing the Balance Controller

In this section, ths state-feedback control that was designed and simulated in the previous sections is run on the
actual SRV02 Rotary Pendulum device.

1. In the Rotary Pendulum (Student) LabVIEW project, open the ROTPEN Control VI. Make sure the HIL Initialize
block has been confirued for your data acquisition device, as explained in Section 5.3.

2. Turn ON the power amplifier.

Ensure the modifications you made to the VI in the Modeling Laboratory (Section 2.3) have been ap-
plied. Run the VI and verify that the model conventions still hold (e.g., motor turns in expected way when a
positive voltage is supplied).

3. As shown in Figure 3.10, the VI block diagram is incomplete. Similarly as done in Section 3.4.2, add the
necessary VIs to implement the balance control. When implementing the control, keep in mind the following:

• Unlike in the simulation, where the pendulum is already upright, the pendulum begins in the hanging down
position. Thus when the VI starts, the inverted pendulum angle reads ±180 and it goes up to zero when
brought to the upright position. You will need to add a switch logic to implement the Equation 3.2.6.

• Recall in the simulation, the Gain VI outputs a vector. Use the first element of the vector to drive the
analog output channel. Make sure you use the proper conventions as dictated in Section 2.3.
• Saturate your control voltage to ±10 V. On initial tests, you may want to set it to a lower range (e.g., ±5V).
4. Go to the front panel of the VI and enter the gain K you found in Section 3.4.1.

ROTPEN Workbook - Student Version 22

 Figure 3.10: Incomplete block diagram of ROTPEN Control Simulation (Student).vi

5. Ensure the pendulum is in the hanging down position and is motionless. Run the VI and manually bring up the
pendulum to its upright, vertical position. You should feel the voltage kick-in when it is within the range where
the balance control engages. Once it is balanced, introduce the ±20 degree rotary arm command by setting
Amplitude (deg) to 20 in the VI. The response should look similar to your simulation. Attach the measured
rotary pendulum responses.
6. Measure the pendulum deflection and voltage used. Are the specifications given in Section 3.1 satisfied for
the implementation?
7. Click on the STOP button and close the VI when done. Be careful, as the pendulum will fall down when the VI
is stopped.
8. Shut off the power amplifier.

ROTPEN Workbook - Student Version v 1.1

3.5 Results
Fill out Table 3.1 with your answers from your control lab results - both simulation and implementation.

Description Symbol Value Units

Pre Lab Questions
Desired poles DP
Companion Gain K̃
Simulation: Control Design

Transformation Matrix W

Control Gain K
Closed-loop poles CLP
Simulation: Closed-Loop System
Maximum deflection |α|max deg
Maximum voltage |Vm |max V
Implementation
Control Gain K
Maximum deflection |α|max deg
Maximum voltage |Vm |max V

Table 3.1: Results

ROTPEN Workbook - Student Version 24

4 SWING-UP CONTROL
4.1 Background
In this section a nonlinear, energy-based control scheme is developed to swing the pendulum up from its hanging,
downward position. The swing-up control described herein is based on the strategy outlined in [7]. Once upright,
the control developed in Section 3 can be used to balance the pendulum in the upright vertical position.

4.1.1 Pendulum Dynamics

The dynamics of the pendulum can be redefined in terms of pivot acceleration as

1 1
Jp α̈ + mp gLp sin(α) = mp Lp u cos(α). (4.1)
2 2

The pivot accleration, u, is the linear acceleration of the pendulum link base. The acceleration is proportional to the
torque of the rotary arm and is expressed as
τ = mr Lr u (4.2)
where mr is the mass of the rotary arm and Lr is its length, as shown in Section 2. The voltage-torque relationship
is given in Equation 2.4.

4.1.2 Energy Control

If the arm angle is kept constant and the pendulum is given an initial position it would swing with constant amplitude.
Because of friction there will be damping in the oscillation. The purpose of energy control is to control the pendulum
in such a way that the friction is constant.

The potential and kinetic energy of the pendulum is

1
Ep = mp gLp (1 − cos(α)) (4.3)
2
and
1
Ek = Jp α̇2 .
2

The pendulum parameters are described in Section 2 and their values are given in [3]. In the potential energy
L
calculation, we assume the center of mass to be in the center of the link, i.e., 2p . Adding the kinetic and potential
energy together give us the total pendulum energy
1 1
E= Jp α̇2 + mp gLp (1 − cos α). (4.4)
2 2

Taking its time derivative we get ( )

1
Ė = α̇ Jp α̈ + mp gLp sin α . (4.5)
2

To introduce the pivot acceleration u and eventually, our control variable, solve for sin α in Equation 4.1 to obtain
1
sin(α) = (−2Jp α̈ + mp Lp u cos(α)).
mp gLp

Substitute this into Ė, found in Equation 4.5, to get

1
Ė = mp Lp uα̇ cos α
2

ROTPEN Workbook - Student Version v 1.1

One strategy that will swing the pendulum to a desired reference energy Er is the proportional control

u = (E − Er )α̇ cos α.

By setting the reference energy to the pendulum potential energy, i.e., Er = Ep , the control will swing the link to its
upright position. Notice that the control law is nonlinear because the proportional gain depends on the pendulum
angle, α, and also notice that the control changes sign when α̇ changes sign and when the angle is ±90 degrees.

For energy to change quickly the magnitude of the control signal must be large. As a result, the following swing-up
controller is implemented
u = satumax (µ(E − Er )sign(α̇ cos α)) (4.6)
where µ is a tunable control gain and satumax function saturates the control signal at the maximum acceleration of
the pendulum pivot, umax . Taking the sign of α̇ cos α allows for faster switching.

In order to translate the pivot acceleration into servo voltage, first solve for the voltage in Equation 2.4 to get
τ Rm
Vm = + Kg km θ̇.
ηg Kg ηm kt

Then substitute the torque-acceleration relationship given in Equation 4.2 to obtain the following
Rm mr Lr u
Vm = + Kg km θ̇. (4.7)
ηg Kg ηm kt

4.1.3 Self-Erecting Control

The energy swing-up control can be combined with the balancing control in Equation 3.11 to obtain a control law
which performs the dual tasks of swinging up the pendulum and balancing it. This can be accomplished by switching
between the two control systems.

Basically the same switching used for the balance control in Equation 3.2.6 is used. Only instead of feeding 0 V
when the balance control is not enabled, the swing-up control is engaged. The controller therefore becomes
{
K(xd − x) |x2 | < ϵ
u= (4.8)
satumax (µ(E − Er )sign(α̇ cos α)) otherwise

ROTPEN Workbook - Student Version 26

4.2 Pre-lab Questions
1. Evaluate the potential energy of the pendulum when it is in the downward and upright positions.

2. Compute the maximum acceleration deliverable by the SRV02. Assume the maximum equivalent voltage
applied to the DC motor is 5 V such that
Vm − Kg km θ̇ = 5. (4.9)
The SRV02 motor parameters are given in [2].

3. Find the controller acceleration when the pendulum is initially hanging down and motionless. From a pracitcal
viewpoint, what does this imply when the swing-up control is activated?

4. Assume the pendulum is starting to swing from the downward position in the positive direction. Calculate the
acceleration the swing-up controller will generate when µ = 20. Does this saturate the controller?

ROTPEN Workbook - Student Version v 1.1

4.3 In-lab Exercises
In this section, you will be implementing the energy-based swing-up controller described in Section 4.1.

1. In the Rotary Pendulum LabVIEW project, open the ROTPEN Swing-Up Control VI. Make sure the HIL Initialize
block in the VI is configured for your data acquisition system, as explained in Section 5.3.

2. Ensure the modifications you made to the VI in the Controls Laboratory in Section 3.4 have been ap-
plied. Run the VI and verify that the balance control runs fine.

3. The incomplete block diagram is shown in Figure 4.1. Edit the Formula Node in the VI to measure the total
energy of the pendulum.

Figure 4.1: Incomplete block diagram in ROTPEN Swing-Up Control VI.

4. Run the VI and rotate the pendulum up to the upright position. While the inverted pendulum is balancing,
record the total energy reading displayed in E (J) numeric indicator. Is the value as expected?
5. Implement the energy-based swing-up controller. Use the LabVIEW controls E_r (J), u_max (m/s2 ), and mu
that are already included. Make sure you are using the full pendulum angle α, i.e., not the upright based angle
used in the feedback for the inverted pendulum balance control.

6. Add the necessary modifications to convert the acceleration generated by the swing-up control to servo voltage.
Use the SRV02 Model Parameters bundle that is already included in the VI for any of the servo-based attributes
you need.

7. Implement the self-erecting control in Equation 4.8, which includes both the swing-up and balance control.
8. Run the VI. Start with the following reference energy, maximum acceleration, and proportional gain parameters:

Er = Ep
umax = 5 m/s2
µ = 2

Make sure the reference energy is set to the pendulum potential energy. The pendulum should be moving back
and forth slowly. Gradually increase the umax and/or µ until the pendulum goes up. Do not increase the umax
above the maximum acceleration you found for the SRV02 in Section 4.2. When the pendulum swings up to
the vertical upright position, the balance controller should engage and balance the link. Show the response
of the arm and pendulum angles as well as the control voltage and record the swing-up parameters. Did the
swing-up behave with the parameters you expected?

ROTPEN Workbook - Student Version 28

 9. Click on the STOP button and close the VI when done. Be careful, as the pendulum will fall down when the VI
is stopped.
10. Shut off the power amplifier.

ROTPEN Workbook - Student Version v 1.1

4.4 Results
Fill out Table 4.1 with your answers from your swing-up control lab results.

Description Symbol Value Unit

Pre Lab Questions
Potential Energy Ep J
Maximum Acceleration of SRV02 umax m/s2
Implementation
Balance Control Gain K
Reference Energy Er J
Control Maximum Acceleration umax m/s2
Swing-Up Proportional Gain µ

Table 4.1: Results

ROTPEN Workbook - Student Version 30

5 SYSTEM REQUIREMENTS
Required Hardware

• Data-aquisition (DAQ) device that is compatible with LabVIEW™ and Quanser Rapid Control Prototyping
Toolkitr .

• Quanser SRV02-ET rotary servo. See Reference [2].

• Quanser Rotary Pendulum Module (attached to SRV02). See Reference [3].

• Quanser VoltPAQ power amplifier, or equivalent (e.g. Reference [4] for VoltPAQ User Manual).

Required Software

• NI LabVIEW™
• NI LabVIEW Control Design and Simulation Module

• NI LabVIEW MathScript RT Module or later

• For NI CompactRIO users:

– NI LabVIEW Real-Time Module
– NI LabVIEW FPGA Module
– NI-RIO Drivers

• Quanser Rapid Control Prototyping Toolkitr

5.1 Overview of Files

File Name Description

Rotary Pendulum Workbook (Stu- This laboratory guide contains pre-lab questions and lab
dent).pdf experiments that involve modeling the pendulum, design-
ing and implementing an inverted pendulum balance con-
trol, and running an energy-based swing-up control.
Rotary Pendulum (Student).lvproj LabVIEW project containing the student-based VIs.
ROTPEN Modeling (Student).vi Creates the inverted pendulum state-space model.
ROTPEN Control Design (Student).vi Design the state-feedback gain through pole-placement
(PP). The model has not been entered (e.g., to be com-
pleted by the student).
ROTPEN Control Simulation (Stu- Simulates the closed-loop response of the rotary pendu-
dent).vi lum. The model has not been enterred and the feedback
loop is not completed.
ROTPEN Control (Student).vi Implements the state-feedback balance controller on
Quanser Rotary Pendulum (ROTPEN). Control loop not
completed.
SRV02 RP Swing Up ((Student).vi Implements the swing-up control on the Quanser Rotary
Pendulum (ROTPEN). Controller not completed.

Table 5.1: Files supplied with the SRV02 Inverted Pendulum laboratory.

ROTPEN Workbook - Student Version v 1.1

5.2 Hardware Setup
Follow these steps to get the system hardware ready for this lab:

1. Make sure the SRV02 is in the high-gear configuration.

2. Install the Rotary Inverted Pendulum module on top of the SRV02 gear as shown in [3].
3. Connect the Quanser Inverted Pendulum to the amplifier (e.g. VoltPAQ) and DAQ device as described in [3].

Note: If you are using the NI CompactRIO, then see the SRV02 cRIO User Manual ([5]).

5.3 Software Setup

Follow these steps to get the system ready for this lab:

1. Load the LabVIEW™ software.

2. Open the LabVIEW project called Rotary Pendulum (Student).lvproj shown in Figure 5.1.

Figure 5.1: Rotary Pendulum LabVIEW Project

3. The Control Design and Simulation folder contains the simulation-based VIs that do not require any hardware.

4. To implement the state-feedback balance controller on the Quanser Rotary Pendulum, open ROTPEN Con-
trol.vi, shown in Figure 5.2.

ROTPEN Workbook - Student Version 32

 Figure 5.2: Front panel of ROTPEN Control (Student).vi

5. To implement the swing-up control on the Quanser Rotary Pendulum, open one the ROTPEN Swing Up.vi,
shown in Figure 5.3.

Figure 5.3: Front panel of ROTPEN Swing Up Control (Student).vi

6. Configure DAQ: Before running the VI, make sure you set the correct Board type (e.g., 'q1_cRIO', 'q2_usb',
'q8_usb', 'qpid', or 'qpid_e') in the HIL Initialize block.
7. Quanser CompactRIO Users: Before running the VI, make sure you can connect to your CompactRIO
through the Measurement & Automation software. See the SRV02 cRIO User Manual ([5]).
8. Channel Configuration: For any of the DAQ-based VIs, the encoder input and ouput channels are set, by
default, to match the wiring in the Rotary Pendulum User Manual ([3]). If the wiring is different on your system,
make sure the VI uses the correct channels. For instance, if your pendulum encoder is connected to Encoder
Input Channel #2 on your DAQ, then set the HIL Encoder Read VI to read from channels [0,2] (instead of [0,1]).

ROTPEN Workbook - Student Version v 1.1

6 LAB REPORT
This laboratory contains three experiments, namely,

1. Modeling,

2. Balance Control, and

3. Swing-Up Control.

When you are writing your lab report, follow the outline corresponding to the experiment you conducted to build the
content of your report. Also, in Section 6.4 you can find some basic tips for the format of your report.

6.1 Template for Content (Modeling)

I. PROCEDURE

1. Model Analysis

• Briefly describe the main goal of the simulation.

• Briefly describe the simulation procedure in steps 2 and 3 in Section 2.3.1.
2. Calibration

• Briefly describe the main goal of the experiment.

• Briefly describe the experiment procedure (Section 2.3.2)

II. RESULTS
Do not interpret or analyze the data in this section. Just provide the results.

1. State-space representation from step 3 in Section 2.3.1.

2. Provide applicable data collected in this laboratory (from Table 2.1).

III. ANALYSIS
Provide details of your calculations (methods used) for analysis for each of the following:

1. Measured arm response in step 7 in Section 2.3.2.

IV. CONCLUSIONS
Interpret your results to arrive at logical conclusions for the following:

1. Whether the arm and pendulum angles match the model conventions in step 4 of Section 2.3.2, Sensor cali-
bration.

2. Whether the control voltage matches the model conventions in step 7 of Section 2.3.2, Actuator calibration.

ROTPEN Workbook - Student Version 34

6.2 Template for Content (Balance Control Experi-
ment)
I. PROCEDURE

1. Control Design
• Briefly describe the main goal of the control design.
• Briefly describe the control design procedure in Step 4 of Section 3.4.1.

2. Simulation
• Briefly describe the main goal of the simulation.
• Briefly describe the simulation procedure (Section 3.4.2).

3. Implementation
• Briefly describe the main goal of this experiment.
• Briefly describe the experimental procedure (Section 3.4.3).

II. RESULTS

Do not interpret or analyze the data in this section. Just provide the results.

1. Matrices from step 4 in Section 3.4.1, Find transformation matrix.

2. Completed block diagram of Pole Placement tab from Step 7 in Section 3.4.1, Block diagram to generate the
control gain.

3. Response plot from step 4 in Section 3.4.2, Inverted pendulum balance control simulation.
4. Response plot from step 5 in Section 3.4.3, Inverted pendulum control implementation.

5. Provide applicable data collected in this laboratory (from Table 3.1).

III. ANALYSIS

Provide details of your calculations (methods used) for analysis for each of the following:

1. Step 3 in Section 3.4.1, Controllability of system.

2. Step 5 in Section 3.4.2, Balance control simulation.

3. Step 6 in Section 3.4.3, Balance control implementation.

IV. CONCLUSIONS

Interpret your results to arrive at logical conclusions for each of the following:

1. Step 6 in Section 3.4.1, Closed-loop system poles.

2. Step 5 in Section 3.4.2, Balance control simulation.

3. Step 6 in Section 3.4.3, Balance control implementation.

ROTPEN Workbook - Student Version v 1.1

6.3 Template for Content (Swing-Up Control Ex-
periment)
I. PROCEDURE

1. Implementation
• Briefly describe the main goal of this experiment.
• Briefly describe the experimental procedure in Step 3 in Section 4.3.
• Briefly describe the control states and parameters for the swing-up control in Step 5 in Section 4.3.
• Briefly describe the parameters used to convert torque to voltage in Step 6 in Section 4.3.

II. RESULTS

Do not interpret or analyze the data in this section. Just provide the results.

1. Completed block diagram in 5 in Section 4.3, Implemented swing-up controller.

2. Completed block diagram in 7 in Section 4.3, Implemented self-erecting controller.

3. Response plot from step 8 in Section 3.4.3, Self-erecting inverted pendulum control implementation response.

4. Provide applicable data collected in this laboratory (from Table 4.1).

III. ANALYSIS

Provide details of your calculations (methods used) for analysis for each of the following:

1. Step 4 in Section 4.3, Potential energy of pendulum.

IV. CONCLUSIONS

Interpret your results to arrive at logical conclusions for each of the following:

1. Step 8 in Section 4.3, Swing-up control implementation.

ROTPEN Workbook - Student Version 36

6.4 Tips for Report Format
PROFESSIONAL APPEARANCE

• Has cover page with all necessary details (title, course, student name(s), etc.)

• Each of the required sections is completed (Procedure, Results, Analysis and Conclusions).
• Typed.

• All grammar/spelling correct.

• Report layout is neat.

• Does not exceed specified maximum page limit, if any.

• Pages are numbered.

• Equations are consecutively numbered.

• Figures are numbered, axes have labels, each figure has a descriptive caption.

• Tables are numbered, they include labels, each table has a descriptive caption.

• Data are presented in a useful format (graphs, numerical, table, charts, diagrams).
• No hand drawn sketches/diagrams.

• References are cited using correct format.

ROTPEN Workbook - Student Version v 1.1

REFERENCES
[1] Bruce Francis. Ece1619 linear systems course notes (university of toronto), 2001.
[2] Quanser Inc. SRV02 User Manual, 2009.

[3] Quanser Inc. SRV02 Rotary Pendulum User Manual, 2010.

[4] Quanser Inc. VoltPAQ User Guide, 2010.

[5] Quanser Inc. SRV02 cRIO User Manual, 2011.

[6] Norman S. Nise. Control Systems Engineering. John Wiley & Sons, Inc., 2008.

[7] K. J. Åström and K. Furuta. Swinging up a pendulum by energy control. 13th IFAC World Congress, 1996.

ROTPEN Workbook - Student Version 38

Over ten rotary experiments for teaching fundamental and advanced controls concepts

Rotary Servo Base Unit Inverted Flexible Link Flexible Joint

Pendulum

Ball and Beam Double Inverted Gyro/Stable 2 DOF Robot

Pendulum Platform

2 DOF Inverted 2 DOF Gantry Multi-DOF Torsion

Pendulum

Quanser’s rotary collection allows you to create experiments of varying complexity – from basic to advanced. Your lab
starts with the Rotary Servo Base Unit and is designed to help engineering educators reach a new level of efficiency
and effectiveness in teaching controls in virtually every engineering discipline including electrical, computer, mechanical,
aerospace, civil, robotics and mechatronics. For more information please contact [email protected]
©2013 Quanser Inc. All rights reserved.

[email protected] +1-905-940-3575 QUANSER.COM

Solutions for teaching and research. Made in Canada.