You are on page 1of 1

Using a Two-Wheeled Balancing Scooter with Variable Configuration to Teach

Control Systems Design.


Jaime Parra-Raad; Sebastian Roa-Prada, ASME Member (Advisor); Hernando Gonzalez (Co-advisor)
Dept. of Mechatronics Engineering, Universidad Autonoma de Bucarmanga, Bucaramanga, Colombia.

Parameters of theplant Time Response System 1


ABSTRACT
Transfer function PID
Parameters System 1 System 2
A two-wheeled balancing scooter is a transportation system inspired by the behavior of the Pendulum mass[Kg] 0.96401 1.73859
inverted pendulum, IP. The reason why the single-degree-of freedom IP system is attractive as a
case study for teaching of control systems design is that its dynamics is non-lineal and unstable, Pendulum radius [m] 0.05 0.015 0.4

Angular position[rads]
requiring a control effort to achieve steady-state equilibrium at a given angular position. An IP Constant torque 0.27468 0.27468 0.3
x: 0

could be balanced at an angle by means of several different types of control action: a fixed rotary generated 0.2
y: 0.3

motor at the pivot, a fixed pivot and a motor at the moving end, the pivot mounted on a moving 0.1

base with rectilinear or 2D curvilinear motion, hydraulic balancing, etc. The architecture of a
Radius of the wheel [m] 0.045 0.045 0
x: 0.1541
y: -0.0228
x: 0.7347
y: -0.01502 Angular position

balancing scooter is then obtained by connecting two independently actuated, concentrically Gravitational acceleration l 9.81 9.81 -0.1
aligned, wheels to the pivot of the pendulum. With this addition, the angular position of the [m/s^2] 0 0.5 1 1.5 2
Time(s)
pendulum is coupled with two other degrees of freedom: translation at the base, and rotation of 0.0003825
Pendulum inertia[kg*m^2] 0.0016082714 15 Control Aaction
the whole structure with respect to the vertical axis. The dynamic response of the balancing x: 0
y: 12
scooter has been fully studied and almost each school of engineering has built a physical full-scale Inertia of the wheel 0.0063 8.525*10^-8 10

Amplitude[v]
model to evaluate the accuracy of the modeling and to test different controllers. However, the [kg*m^2] 5
x: 0.1541 x: 0.7347
usual balancing scooters built do not allow for changes in geometry and configuration. This y: -0.1599 y: -0.2033

research proposes a flexible configuration balancing scooter, in which several constructional


Wheel mass [kg] 0.14 0.0148 0

parameters can be modified, such as the size of the wheels, total weight, position of the center of Inductance in the rotor 0.00143 0.00143 -5
0 0.5 1 1.5 2
Time(s)
mass, etc., in order to explore the ability of a control strategy to respond to changes in the plant. [H]
Different types of control strategies can be implemented such as lead-lag, PID and state space Constant voltage generated 0.7445 0.7445
controllers, etc. With the approach proposed, the students can have the opportunity to observe Time Response System 2
changes in the system’s response with respect to changes in the plant for a given type of controller, Motor armature resistance 11.1 11.1
adjusting the parameters of the controller to maintain balance of the system at all times. Similarly, [ohm]
the students can assess the efficiency of a certain type of controller to give the plant the desired
Transfer function PID
performance, and the ability of the actuators to stabilize the system before reaching saturation, as
well as the effects of the dynamic response of the sensors on the overall dynamic response of the
controlled system. This approach will help introducing undergraduate students to more advanced
techniques such as adaptive and optimal control. Linearization x: 0
0.3

Angular position[rads]
y: 0.3
Angular position
0.2
0.1 x: 0.2823 x: 1.069
y: 0.02034 y: -4.037e-006
0
System 1
INTRODUCTION 0.1
0.2
x: 0.1209
y: -0.1348

0 0.5 1 1.5 2
Time(s)
20 x: 0.0009097 Control action
The balancing scooter system under study is in essence inverted pendulum on wheels. The first + +
10
y: 10.38

Amplitude[v]
step in this analysis is the derivation of the correspondingequations of motion of the scooter,
1
0
which will permitpredictingits dynamic behavior. The physical system consists of two wheels, x: 0.2139
y: -0.0006301

around which the pendulum system hinges; and , two DC motors as the actuators of the system. A -10 x: 0
y: -12
schematic of the pendulum is shown in the figure below. -20
System 2 0 0.5 1 1.5 2
Time(s)

+ +
DISCUSSION AND SUMMARY
The control action is generated and transmited to the pendulum by changing the
angular and linear velocity of the wheels, which can only occur thanks to the
torque delivered by the motor.

CRONTOL SYSTEM DESIG N When the distance between the center of mass and the pivot is reduced, two
features can be distinguished in the response of the system: first the pendulum
will tend to fall more slowly and then the plant will change from having negative
frequency responses to positive ones. The relation with constants PID is when the
plant have negative frequency responses, the PID constants are negative, while it
Discrete Parallel PID Controler will be positive when the plant have positive frequency responses.

MODELING By reducing the moment of iner tia of the pendulum with respect to the pivot, we
see that the plant changes from having two real poles, one on the negative real
axis and the other on the positive real axis, creating a root locus in the positive
axis, to two complex poles in the first and four th quadrant, which facilitates the
design of the PID structure since it made the change between the locus of the
Free Body Diagrams roots to the second and third quadrant easier, giving stability to the plant.

The plant can be stabilized with less control effor t, providing the required angular

mrθ^2
Tf
RESULTS acceleration with the same actuar tors.

R L Ia(t)

mrθ


r
Oy
Td

+x Ta(t) Frequency Response


REFERENCES
mwx Ox Va(t)
mx
mg Iwx/r mwg
Tm Ff Tm(t)
Tf C System 1 Impulse Response [1]
θ(t) Time (seconds): 0 System 2

Marion, Jerr y B. (1996) (en español). “Dinámica clásica de las par tículas y
Time (seconds): 4.37
Ox ω(t)
Amplitude: 0
System 1
Amplitude: 5.04e+029

Time (seconds): 2.14


N Amplitude: -4.19e+012
sistemas”,Ed.Rever té,Barcelona. ISBN 84-291-4094-8.
Oy System 2
[2]
System 2 System 2
Time (seconds): 2.14 Time (seconds): 4.42
Time (seconds): 0 Amplitude:-2.66e+028
Amplitude: 5.63e+014
Amplitude: 0 System 1
Time (seconds):4.37
Amplitude:-1.54e+026 Ferdinand P. Beer; E.Russell Johnston, JR y Phillip J. Cornwell .2010. “Mecánica vectorial para
ingenieros dinámica. Mc Graw Hill. ISBN-13:978-607-15-0261-2
Equations of motion
System 1
Time (seconds):4.42
Amplitude:-3.14e+026 [3]
W. Kaplan,1985, “Matemáticas avanzadas para estudiantes de ingeniería,México”, Fondo
Amplitude

Educativo Interamericano.
[4] ,
θ θ θ J. Zabcz yk,1992.“Mathematical control theor y : An introduction”,Ed. Birkhäuser,Boston.
[5]
x θ θ θ θ
Dennis G. Zill. 2009. “Ecuaciones diferenciales con aplicaciones al modelado”. CENGAGE
Learning. ISBN-13:978-970-830-055-1
System 1
Sytem 2 [6]
Katsuhiko Ogata. 2010. “Modern control engineering”.pearson. ISBN 13: 978-0-13-615673-4
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Time(seconds)

You might also like