World Academy of Science, Engineering and Technology

International Journal of Mechanical and Mechatronics Engineering

Vol:4, No:10, 2010

The Control of a Highly Nonlinear Two-wheels

Balancing Robot: A Comparative Assessment
between LQR and PID-PID Control Schemes
A. N. K. Nasir, M. A. Ahmad, R. M. T. Raja Ismail

Due to these reasons, increasing effort has been made towards

Abstract—The research on two-wheels balancing robot has control designs that guarantee stability and robustness for
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

gained momentum due to their functionality and reliability when mobile wheeled inverted pendulums. Although two wheels
completing certain tasks. This paper presents investigations into the balancing robot are intrinsically nonlinear and their dynamics
performance comparison of Linear Quadratic Regulator (LQR) and
will be described by nonlinear differential equations, it is
PID-PID controllers for a highly nonlinear 2–wheels balancing robot.
The mathematical model of 2-wheels balancing robot that is highly often possible to obtain a linearized model of the system. If
nonlinear is derived. The final model is then represented in state- the system operates around an operating point, and the signals
space form and the system suffers from mismatched condition. Two involved are small signals, a linear model that approximates
system responses namely the robot position and robot angular the nonlinear system in the region of operation can be
position are obtained. The performances of the LQR and PID-PID obtained. Several techniques for the design of controllers and
controllers are examined in terms of input tracking and disturbances
analysis techniques for linear systems were applied. In [2],
rejection capability. Simulation results of the responses of the
nonlinear 2–wheels balancing robot are presented in time domain. A motion control was proposed using linear state-space model.
comparative assessment of both control schemes to the system In [3], dynamics was derived using a Newtonian approach and
performance is presented and discussed. the control was design by the equations linearized around an
operating point. In [4], the dynamic equations were studied,
Keywords—PID, LQR, Two-wheels balancing robot. with the balancing robot pitch and the rotation angles of the
two wheels as the variables of interest, and a linear controller
I. INTRODUCTION was designed for stabilization under the consider of its

T HE research on two-wheeled balancing robot has gained

momentum over the last decade due to the nonlinear and
unstable dynamics system. Various control strategies had been
robustness in [5]. In [6], a linear stabilizing controller was
derived by a planar model without considering vehicle yaw.
The above control laws are designed on the linearized
proposed by numerous researchers to control the two-wheeled dynamics which only exhibits desirable behavior around the
balancing robot such that the robot able to balance itself. Two operating point, and do not have global applicability. In [7],
wheels balancing robot is a good platform for researchers the exact dynamics of two wheels inverted pendulum was
to investigate the efficiency of various controllers in investigated, and linear feedback control was developed on
control system. The research on two wheels balancing robot is the dynamic model. In [8], a two-level velocity controller via
based on inverted pendulum model. Thus, a two wheels partial feedback linearized and a stabilizing position controller
balancing robot needs a good controller to control itself in were derived; however, the controller design is not robust
upright position without the needs from outside. with respect to parameter uncertainties. In [9], a controller
Motion of two wheels balancing robot is governed by using sliding mode approach was proposed to ensure
under-actuated configuration, i.e., the number of control robustness versus parameter uncertainties for controlling both
inputs is less than the number of degrees of freedom to be the position and the orientation of the balancing robot.
stabilized [1], which makes it difficult to apply the The mathematical model is established through a modeling
conventional robotics approach for controlling the systems. process where the system is identified based on the
conservation laws and property laws. This process is crucial
since a controller is design solely based on this mathematical
A.N.K. Nasir is with Control & Instrumentation Research Group (COINS), model. Thus, an accurate equation must be derived in order
Faculty of Electrical & Electronics Engineering, Universiti Malaysia Pahang,
for the controller to response accordingly.
Pekan, Pahang, Malaysia (phone: 609-4242145-; fax: 609-424-2035; e-mail:
kasruddin@ ump.edu.my).
M.A. Ahmad is with Faculty of Electrical & Electronics Engineering,
Universiti Malaysia Pahang, Pekan, Pahang, Malaysia (e-mail:
mashraf@ump.edu.my).
R.M.T. Raja Ismail is with Faculty of Electrical & Electronics Engineering,
Universiti Malaysia Pahang, Pekan, Pahang, Malaysia (e-mail:
rajamohd@ump.edu.my).

International Scholarly and Scientific Research & Innovation 4(10) 2010 942 scholar.waset.org/1307-6892/15128
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:4, No:10, 2010

This paper presents investigations of performance TABLE I

LIST OF PARAMETERS OF TWO-WHEELS BALANCING ROBOT BASED ON [11]
comparison between conventional (PID-PID) and modern
control (Linear Quadratic Regulator) schemes for a two Symbol Parameter Value
wheels balancing robot. The mathematical model of the two
D distance between 0.2 m
wheels balancing robot system is presented in differential contact patches of
equation form. The dynamic model of the system with the the wheels
permanent magnet DC motors dynamic included is derived g gravitational force 9.81 m.s-2
Jp chassis’s inertia 0.0041 kg.m2
based on [3] and [10]. Performances of both control strategy Jpδ chassis’s inertia during 0.00018 kg.m2
with respect to balancing robot outputs angular position θ and rotation
linear position x are examined. Comparative assessment of Jw wheel’s inertia 0.000039 kg.m2
ke back emf constant 0.006087 Vs/rad
both control schemes to the two balancing robot system km motor torque constant 0.006123 Nm/A
performance is presented and discussed. l distance between 0.07 m
center of the wheels
II. DYNAMIC MODEL and the robot’s CG
Mp body’s mass 1.13 kg
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

Modeling is the process of identifying the principal Mw wheel’s mass 0.03 kg

physical dynamic effects to be considered in analyzing a R nominal terminal 3Ω
resistance
system, writing the differential and algebraic equations r wheel’s radius 0.051 m
from the conservative laws and property laws of the
relevant discipline, and reducing the equations to a convenient
differential equation model. This section provides a
description on the modeling of the two wheels balancing
robot, as a basis of a simulation environment for development
and assessment of both control schemes. The robot with its
three degrees of freedom is able to linearly move which is
characterized by position x, able to rotate around the y-axis
(yaw) with associated angle δ and able to rotate around z-axis
(pitch) where the movement is described by angle θ. List of
parameters for the two wheels balancing robot are shown in
Table I. These parameters are based on the project conducted
by Ooi (2003) as stated by [10]. The inputs of the system are
the voltages VaR and VaL which both are applied to the two
motors which located on right side and left side of the robot as
shown in Fig. 1. In order to obtain the dynamic model of the
Fig. 1 A mobile balancing robot (Grasser et al., 2002)
balancing robot some assumptions and limitations are
introduced:
1) Motor inductance and friction on the motor armature is
neglected.
2) The wheels of the robot will always stay in contact with
the ground.
3) There is no slip at the wheels.
4) Cornering forces are also negligible.

Fig. 2 shows a free body diagram of the balancing robot

which contributed to the nonlinear dynamic equations of the
system. Equation (1) represents linear acceleration in x
direction, equation (2) represents angular acceleration about y-
Fig. 2 Free body diagram of balancing robot
axis and equation (3) represents angular acceleration about z-
axis.
As can be seen from equations (1), (2), and (3), all
nonlinear terms are remain in the equations. All these
equations are used to design the proposed controllers which
will be described in the section III.

International Scholarly and Scientific Research & Innovation 4(10) 2010 943 scholar.waset.org/1307-6892/15128
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:4, No:10, 2010

2 k e k m ⎡ 1 M p l cos θ ⎤ A. PID Controller

&x& = − ⎢ + ⎥ x&
α rR β ⎣ r γ ⎦ PID stands for Proportional-Integral-Derivative. This is a
M 2p gl 2 sin θ cos θ type of feedback controller whose output, a control variable
+ θ (CV), is generally based on the error (e) between defined set
αγβθ
point (SP) and some measured process variable (PV). Each
k m ⎡ 1 M p l cos θ ⎤
+ ⎢ + ⎥V aR element of the PID controller refers to a particular action
αRβ ⎣ r γ ⎦
taken on the error. In order to demonstrate the performance of
k m ⎡ 1 M p l cos θ ⎤ 1 the PID controller in locating the balancing robot to its
+ ⎢ + ⎥V aL + f drR
αRβ ⎣ r γ ⎦ αβ (1) desired position and angle, the collocated sensor signal of the
1 1 ⎡ M p l 2 cos 2 θ ⎤ position of the robot about roll axis, x(s) and angular position
+ f drL + ⎢1 + ⎥ f dp of the robot about yaw axis θ(s) are fed back and compared to
αβ αβ ⎢⎣ γ ⎥⎦
the reference position, xf(s) and angle θf(s) respectively.
& 2
M p l θ sin θ
+ . Initially, the angular position of the robot which is position
αβ about pitch axis is set 50 degrees or 0.8727 radians. In this
study, two PID controllers are required to control the position
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

2kekm ⎡ M p l cos θ ⎤ on the roll axis and the angular position about the yaw axis.
θ&& = ⎢1 + ⎥ x&
γ rR β ⎣ αr ⎦ The position and angular position errors are regulated through
M p gl sin θ M p l cos θ ⎤
the proportional, integral and derivative gain for each PID.
km ⎡
+ θ− ⎢1 + ⎥V aR Block diagram of the PID controller is shown in Fig. 3, where
γβθ γR β ⎣ αr ⎦
u1(s) and u2(s) represent the applied voltage at the right motor
km ⎡ M p l cos θ ⎤ M p l cos θ and left motor respectively. Both of the inputs of the
− ⎢1 + ⎥V aL − f drR
γR β ⎣ αr ⎦ αγβ balancing robot are limited to 20 volts to –20 volts. The
M p l cos θ l cos θ ⎡ Mp⎤ (2) control signal u1(s) and u2(s) in Fig. 3 can be represented as in
+ f drL + ⎢1 − ⎥ f dp
αγβ γβ ⎣ α ⎦ equations (7) and (8) respectively:
M 2p l 2θ& 2 sin θ cos θ ⎛ K ⎞
− . u PID ( s ) position = −⎜ K P1 + I 1 + K D1 s ⎟[r ( s ) − r f ( s )] (7)
αγβ ⎝ s ⎠

km D km D ⎛ K ⎞
δ&& = − V aR + V aL
2 J p δ rR 2 J p δ rR u PID ( s ) angle = −⎜ K P 2 + I 2 + K D 2 s ⎟[θ ( s) − θ f ( s )] (8)
⎝ s ⎠
D D
− f drR + f drL . (3)
2 J pδ 2 J pδ
where s is the Laplace variable. Hence the closed-loop transfer
function is obtained as in equation (9) and (10).
The symbols of α, β, and γ in equations (1), (2), and (3) are
defined as in equation (4), (5), and (6):
⎛ K ⎞
⎜ K P1 + K D1 s + I 1 ⎟G ( s )
r (s) s ⎠
2J w = ⎝ (9)
α = 2M w + +Mp r f (s) ⎛ K (4)⎞
r2 1 + ⎜ K P1 + K D1 s + I 1 ⎟G ( s )
⎝ s ⎠
αγ − M p2 l 2 cos 2 θ
β= (5)
αγ
⎛ KI2 ⎞
γ = J p + M pl 2 (6) ⎜K + K D2 s + ⎟G ( s )
θ (s) P2
s ⎠
= ⎝ (10)
θ f (s) ⎛ K ⎞
III. CONTROLLER DESIGN & SIMULATION 1 + ⎜ K P 2 + K D 2 s + I 2 ⎟G ( s )
⎝ s ⎠
In this section, two control schemes (LQR and PID) are
proposed and described in detail. Furthermore, the following In this paper, the Ziegler-Nichols approach is utilized to
design requirements have been made to examine the design both PID controllers. Analyses the tuning process of
performance of both control strategies. the proportional, integral and derivative gains using Ziegler-
1) The system overshoot (%OS) of robot position, x is to be Nichols technique shows that the optimum response of PID
at most 25%. controller for controlling linear position is achieved by setting
2) The Rise time (Tr) of robot position, x less than 5 s. KP1 = -8, KI1 = -0.921 and KD1 = -6, while for controlling
3) The settling time (Ts) of robot position, x and robot angle angular position, KP2 = -63, KI2 = -60 and KD2 = -11. All the
θ is to be less than 10 seconds. PID1 and PID2 controller parameters must be tuned
4) Steady-state error is within 2% of the initial value. simultaneously to achieve the best responses as desired.

International Scholarly and Scientific Research & Innovation 4(10) 2010 944 scholar.waset.org/1307-6892/15128
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:4, No:10, 2010

∫ x(t )
T
xf(s) u1(s) Nonlinear x(s) J= Qx(t ) + (− Kx(t )) T R( − Kx(t ))dt (13)
PID1 two 0
+ wheels Equation (13) can be further simplified as represented
balancing equation (14).
θf(s) robot ∞

∫ x(t )
T
PID2 θ(s) J= (Q + K T RK ) x(t )dt (14)
+ u2(s) 0
In designing LQR controller, LQR function in matlab m-
file can be used to determine the value of the vector K which
Fig. 3 Block diagram of PID controller
determines the feedback control law. This is done by choosing
two parameter values, input R and Q=C'xC where C is from
state equation of the linearized model. The controller can be
B. LQR Controller tuned by changing the nonzero elements in Q matrix which is
LQR is a method in modern control theory that uses state- done in m-file code. Consequently, by tuning the values of
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

space approach to analyze such a system. Using state-space nonzero elements in matrix Q as shown in (15) and matrix R
methods it is relatively simple to work with a multi-output as shown in (16), the values of matrix K are obtained as
system. The system can be stabilized using full state feedback. represented in equation (17).
The schematic of this type of control system is shown in Fig.
4. Q = [1 0 0 0 0 0; 0 1 0 0 0 0; 0 0 100 0 0 0;
rf(s) + u(s) Nonlinear r(s)
N two 0 0 0 100 0 0; 0 0 0 0 0.1 0; 0 0 0 0 0 0.01] (15)
wheels
_ balancing R = [1 0; 0 1] (16)
robot

The matrix R and matrix Q must be properly tuned to

minimized the quadratic cost function as represented in (14).

K K=[–5.9000 –6.8040 –34.1774 –7.9286 –1.7071 –8.0903;

Fig. 4 The LQR control structure –5.9000 –6.8040 –34.1774 –7.9286 –0.7071 –7.0903] (17)

For a LTI system, the technique involves choosing a control The matrix K is the LQR controller parameter which
law u = ψ (x) which stabilizes the origin while minimizing determines feedback control law u = ψ ( x) = Kx .
the quadratic cost function which can be presented as in
equation (11). IV. RESULT AND ANALYSIS
In this section, the simulation results of the proposed
∞ controller, which is performed on the model of a two wheels
∫
J = x(t ) T Qx(t ) + u (t ) T Ru (t )dt
0
(11) balancing robot are presented. Comparative assessment of
both control strategies to the system performance are also
where Q = Q T ≥ 0 and R = R T > 0 . The term “linear- discussed in details in this section.
Two wheels balancing robot systems with LQR and PID
quadratic” refers to the linear system dynamics and the
controller block diagram produced two responses, angular
quadratic cost function. A famous and somewhat surprising
position θ and linear position x. As stated earlier, the initial
result due to Kalman is that the control law which minimizes
value of the angular position θ of the balancing robot was set
J always takes the form u = ψ ( x) = Kx . The optimal
to 0.5 radians. It means that the initial condition of the robot is
regulator for a LTI system with respect to the quadratic cost very unstable. Fig. 5 shows the comparison of the balancing
function above is always a linear control law. With this robot linear position response between LQR and PID
observation in mind, the closed-loop system takes the states controller graphically. In this figure, the response for the
space form as represented in equation (12). linear position of the robot with PID controller is represented
by straight line or blue color line and the response for the
x& = ( A − BK ) x (12) linear position of the robot with LQR controller is represented
Substituting the control law into equation (11), the cost by dotted line or red color line. Fig. 5 shows that both of the
function J is represented as in equation (13). controllers are capable to control the linear position of the
nonlinear two wheels balancing robot.

International Scholarly and Scientific Research & Innovation 4(10) 2010 945 scholar.waset.org/1307-6892/15128
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:4, No:10, 2010

Balancing Robot Linear Position Responses controller angular position responses. It shows that the result
0.8
has got similar pattern and not much different.
PID
0.7
LQR
Balancing Robot Angular (yaw) Position Rsponses
0.6 0.5
Linear Position (meter)

0.5 PID
0.4
LQR
0.4 0.3

Angular Position (radians)

0.3 0.2
0.2 0.1
0.1 0
0 -0.1
-0.1
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

-0.2
0 1 2 3 4 5 6 7
Time (second) -0.3
Fig. 5 Two-Wheel Balancing Robot Linear Position Response
-0.4

Table II shows the summary of the performance -0.5

characteristics of the balancing robot linear position between 0 1 2 3 4 5
Time (second)
LQR and PID controller quantitatively. Based on the data
tabulated in Table II, LQR has the fastest settling time of 2.38 Fig. 6 Two-Wheels Balancing Robot Angular Position
seconds while PID has the slowest settling time of 2.68 Response
seconds. An extra of 0.3 seconds is required for the PID
controller balancing robot to balance itself. Similarly, for the The initial value of the balancing robot angular position is
maximum overshoot, LQR controller has the best overshoot 0.5 radians. The robot needs to balance itself by eliminating
which is the lowest overshoot between two controllers. The the angular position so that the body of the robot remains
maximum displacement of the balancing robot when LQR vertically straight in upright position. Fig. 6 shows that both
control signal applied to the system is 0.73 meters while of the LQR and PID controllers are capable of controlling the
maximum displacement of the balancing robot when PID nonlinear unstable balancing robot.
control signal applied to the system is 0.77 meters. A distance Table III shows the summary of the performance
of minimum 0.04 meters is required for the PID controller characteristics of the balancing robot angular position between
balancing robot to balance itself. Despite the large initial LQR and PID controller quantitatively. Based on the data
values for the displacement, the proposed LQR controller is tabulated in Table III, PID has the fastest settling time of 2.45
able to bring itself to the vertical position. In term of the rise seconds while LQR has the slowest settling time of 2.59
time, balancing robot with PID controller has the fastest rise seconds. An extra time of 0.14 seconds is required for the PID
time 0.37 seconds while balancing robot with LQR controller controller balancing robot to balance itself. In contrast, for the
needs an extra time of 0.02 seconds to rise from 10% to the maximum undershoot, LQR controller has the best undershoot
90% of the initial value. In term of steady state error, both of which is the lowest undershoot between two controllers. The
the controllers had shown very outstanding performance by maximum angular displacement of the balancing robot when
giving zero error at time 6 seconds and more. The responses LQR control signal applied to the system is –0.29 radians
of the balancing robot linear position have acceptable while maximum angular displacement of the balancing robot
overshoot and undershoot. when PID control signal applied to the system is –0.38
Fig. 6 shows the balancing robot with LQR and PID radians. An extra angle of minimum 0.01 meters is required
TABLE II TABLE III
SUMMARY OF PERFORMANCE CHARACTERISTICS OF THE BALANCING SUMMARY OF PERFORMANCE CHARACTERISTICS OF THE BALANCING
ROBOT LINEAR POSITION BETWEEN LQR AND PID ROBOT ANGULAR POSITION BETWEEN LQR AND PID

Time Response Time Response

LQR PID LQR PID
Spesification Spesification
Rise Time 0.39 sec 0.37sec Rise Time 0.22 sec 0.26sec

Settling Time 2.38 sec 2.68 sec Settling Time 2.59 sec 2.45 sec

Steady state error 0.00 0.00 Steady state error 0.00 0.00

Maximum overshoot 0.73meter 0.77 meter Maximum undershoot 0.29radians 0.38 radians

International Scholarly and Scientific Research & Innovation 4(10) 2010 946 scholar.waset.org/1307-6892/15128
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:4, No:10, 2010

for the PID controller balancing robot to balance itself. V. CONCLUSION

Despite the large initial values for the displacement, the In this paper, two controllers such as LQR and PID are
proposed LQR controller is able to bring itself to the vertical successfully designed. Based on the results and the analysis, a
position. In term of the rise time, balancing robot with LQR conclusion has been made that both of the control method,
controller has the fastest rise time 0.22 seconds while modern controller (LQR) and conventional controller (PID)
balancing robot with PID controller needs an extra time of are capable of controlling the nonlinear two wheels balancing
0.04 seconds to rise from 10% to the 90% of the initial value. robot angular and linear position. All the successfully
In term of steady state error, both of the controllers had shown designed controllers were compared. The responses of each
very outstanding performance by giving zero error at time 4 controller were plotted in one window and are summarized in
seconds and more. Table II and Table III. Simulation results and bar charts in
The responses of the balancing robot angular position have Fig. 7 and Fig. 8 show that LQR controller has better
acceptable overshoot and undershoot. Performance performance compared to PID controller in controlling the
characteristics for linear and angular position represented in nonlinear balancing robot system. Further improvement need
bar chart form are shown in Fig. 7 and Fig. 8 respectively. to be done for both of the controllers. PID controller should be
International Science Index, Mechanical and Mechatronics Engineering Vol:4, No:10, 2010 waset.org/Publication/15128

improved so that the maximum overshoot and maximum

undershoot for the linear and angular positions do not have
3
2.68 very high range as required by the design criteria. On the
2.5 2.38 other side, LQR controller can be improved so that it’s settling
time for angular position might be reduced as faster as PID
2 controller.

1.5 REFERENCES
[1] A. Isidori, L. Marconi, A. Serrani, Robust Autonomous Guidance: An
1 0.77 Internal Model Approach. New York: Springer Verlag, 2003.
0.73
[2] Y.-S. Ha and S. Yuta, “Trajectory tracking control for navigation of the
0.39 0.37 inverse pendulum type self-contained mobile robot,” J. Robotics and
0.5
Autonomous System, vol. 17(1-2), pp. 65–80, Apr. 1996.
[3] F. Grasser, A. Arrigo, S. Colombi and A. C. Rufer, “JOE: a mobile
0 inverted pendulum,” IEEE Trans. Industrial Electronics, vol. 49(1), pp.
Tr Ts Max. OS 107–114, Feb. 2002.
[4] A. Salerno and J. Angeles, “On the nonlinear controllability of a
quasiholonomic mobile robot,” Proc. of IEEE Int.. Conf. on Robotics
LQR PID and Automation, vol. 3, pp. 3379–3384, Sept. 2003.
[5] A. Salerno and J. Angeles, “The control of semi-autonomous two-
Fig. 7 Performance characteristics for linear position wheeled robots undergoing large payload-variations,” Proc. of IEEE Int.
Conf. on Robotics and Automation, vol. 2, pp. 1740-1745, Apr. 2004.
[6] A. Blankespoor and R. Roemer, “Experimental verification of the
3
dynamic model for a quarter size self-balancing wheelchair,” Proc. of
2.59 American Control Conf., pp. 488–492, 2004.
2.45 [7] S. S. Ge and C. Wang, “Adaptive neural control of uncertain MIMO
2.5
nonlinear systems,” IEEE Trans. Neural Networks, vol. 15(3), pp. 674–
692, 2004.
2 [8] K. Pathak, J. Franch and S. K. Agrawal, “Velocity and position control
of a wheeled inverted pendulum by partial feedback linearization,” IEEE
1.5 Trans. Robotics, vol. 21(3), pp. 505–513, 2005.
[9] D. S. Nasrallah, H. Michalska and J. Angeles, “Controllability and
1 posture control of a wheeled pendulum moving on an inclined plane,”
IEEE Trans. Robotics, vol. 23(3), pp. 564–577, 2007.
0.5 0.38 [10] R. C. Ooi, “Balancing a Two-Wheeled Autonomous Robot.” B.Sc. Final
0.22 0.26 0.29
Year Project, University of Western Australia School of Mechanical
0 Engineering, 2003.
Tr Ts Max. US

LQR PID

Fig. 8 Performance characteristics for angular position

International Scholarly and Scientific Research & Innovation 4(10) 2010 947 scholar.waset.org/1307-6892/15128