11

Comparative Analysis of Mobile Robot

Localization Methods Based On Proprioceptive
and Exteroceptive Sensors
Gianluca Ippoliti, Leopoldo Jetto, Sauro Longhi, Andrea Monteri
Dipartimento di Ingegneria Informatica Gestionale e dellAutomazione
Universit Politecnica delle Marche
Via Brecce Bianche 60131 Ancona
Italy

Open Access Database www.i-techonline.com

1. Introduction
Two different approaches to the mobile robot localization problem exist: relative and
absolute. The first one is based on the data provided by sensors measuring the dynamics of
variables internal to the vehicle; absolute localization requires sensors measuring some
parameters of the environment in which the robot is operating. If the environment is only
partially known, the construction of appropriate ambient maps is also required. The actual
trend is to exploit the complementary nature of these two kinds of sensorial information to
improve the precision of the localization procedure (see e.g. (Bemporad et al., 2000; Bonci et
al., 2004; Borenstein et al., 1997; Durrant-Whyte, 1988; Gu et al., 2002; Ippoliti et al., 2004)) at
expense of an increased cost and computational complexity. The aim is to improve the
mobile robot autonomy by enhancing its capability of localization with respect to the
surrounding environment.
In this framework the research interests have been focused on multi-sensor systems because
of the limitations inherent any single sensory device that can only supply partial
information on the environment, thus limiting the ability of the robot to localize itself. The
methods and algorithms proposed in the literature for an efficient integration of multiplesensor information differ according to the a priori information on the environment, which
may be almost known and static, or almost unknown and dynamic.
In this chapter both relative and absolute approaches of mobile robot localization are
investigated and compared. With reference to relative localization, the purpose of this
chapter is to propose and to compare three different algorithms for the mobile robot
localization only using internal sensors like odometers and gyroscopes. The measurement
systems for mobile robot localization only based on relative or dead-reckoning methods,
such as encoders and gyroscopes, have the considerable advantage of being totally selfcontained inside the robot, relatively simple to use and able to guarantee a high data rate. A
drawback of these systems is that they integrate the relative increments and the localization
errors may considerably grow over time if appropriate sensor-fusion algorithms are not
used (De Cecco, 2003). Here, different methods are analysed and tested. The best
performance has been obtained in the stochastic framework where the localization problem
has been formulated as a state estimation problem and the Extended Kalman Filtering (EKF)
Source: Mobile Robots: Perception & Navigation, Book edited by: Sascha Kolski, ISBN 3-86611-283-1, pp. 704, February 2007, Plv/ARS, Germany

www.intechopen.com

216

Mobile Robots, Perception & Navigation

is used. The EKF fuses together odometric and gyroscopic data. A difference with respect to
other EKF based techniques is that the approach followed here derives the dynamical
equation of the state-space form from the kinematic model of the robot, while the measure
equation is derived from the numerical integration equations of the encoder increments.
This allows to fuse together all the available informative content which is carried both by the
robot dynamics and by the acquired measures.
As previously mentioned, any relative localization algorithm is affected by a continuous
growth in the integrated measurement error. This inconvenience can be reduced by
periodically correcting the internal measures with the data provided by absolute sensors
like sonar, laser, GPS, vision systems (Jarvis, 1992; Talluri & Aggarwal, 1992; Zhuang &
Tranquilla, 1995; Mar & Leu, 1996; Arras et al., 2000; Yi et al., 2000; Panzieri et al., 2002). To
this purpose, a localization algorithm based on a measure apparatus composed of a set of
proprioceptive and exteroceptive sensors, is here proposed and evaluated. The fusion of
internal and external sensor data is again realized through a suitably defined EKF driven by
encoder, gyroscope and laser measures.
The developed algorithms provide efficient solutions to the localization problem, where
their appealing features are:

The possibility of collecting all the available information and uncertainties of a

different kind into a meaningful state-space representation,

The recursive structure of the solution,

The modest computational effort.

Significant experimental results of all proposed algorithms are presented here, and their
comparison concludes this chapter.

2. The sensors equipment

In this section the considered sensor devices are introduced and characterized.
2.1 Odometric measures
Consider a unicycle-like mobile robot with two driving wheels, mounted on the left and right
sides of the robot, with their common axis passing through the center of the robot (see Fig. 1).

Fig. 1. The scheme of the unicycle robot.

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

217

Localization of this mobile robot in a two-dimensional space requires the knowledge of

coordinates x and y of the midpoint between the two driving wheels and of the angle
between the main axis of the robot and the X -direction. The kinematic model of the
unicycle robot is described by the following equations:
(1)
x& ( t ) = ( t ) cos ( t )

y& ( t ) = ( t ) sin ( t )

(2)

& ( t ) = ( t )

(3)

where ( t ) and ( t ) are, respectively, the displacement velocity and the angular velocity
of the robot and are expressed by:

r ( t ) + l ( t )

(4)
r
2
( t ) l ( t )
(5)
r
(t ) = r
d
where r ( t ) and l ( t ) are the angular velocities of the right and left wheels, respectively,

(t ) =

is the wheel radius and d s the distance between the wheels.

Assuming constant r ( t ) and l ( t ) over a sufficiently small sampling period tk := tk +1 tk ,

the position and orientation of the robot at time instant

x ( tk +1 ) = x ( tk ) + ( tk ) tk

sin

( tk ) tk
2

( tk ) tk
2

y ( tk +1 ) = y ( tk ) + ( tk ) tk

sin

( tk ) tk
2

( tk ) tk
2

tk +1

can be expressed as:

( tk ) tk

cos ( tk ) +

(6)

( tk ) tk

sin ( tk ) +

(7)

( tk +1 ) = ( tk ) + ( tk ) tk

(8)

where ( tk ) tk and ( tk ) tk are:

qr ( tk ) + ql ( tk )
(9)
r
2
q ( t ) ql ( tk )
(10)
r.
( t k ) t k = r k
d
are the incremental distances covered on the interval tk by

( tk ) tk =

The terms qr ( tk ) and ql ( tk )

the right and left wheels of the robot respectively. Denote by yr ( tk ) and yl ( tk ) the measures
of qr ( tk ) and ql ( tk ) respectively, provided by the encoders attached to wheels, one has
yr ( tk ) = qr ( tk ) + sr ( tk )

(11)

yl ( tk ) = ql ( tk ) + sl ( tk )

(12)

where sr ( ) and sl () are the measurement errors, which are modelled as independent,

www.intechopen.com

218

Mobile Robots, Perception & Navigation

zero mean, gaussian white sequences

follows that the really available values

y ( tk )

and

y ( tk )

(Wang, 1988). It
of ( tk ) tk and ( tk ) tk

respectively are given by:

y r ( tk ) + y l ( tk )
(13)
r = ( tk ) tk + ( tk )
2
y ( t ) y l ( tk )
(14)
y ( tk ) = r k
r = ( tk ) tk + ( tk )
2
and ( ) are independent, zero mean, gaussian white sequences
y ( tk ) =

where ( )

= ( +
2

2
r

2
l

)r

where,

by

(9)

and

(10),

2 = ( r2 + l2 ) r 2 4

and

d .
2

2.2 The Fiber optic gyroscope measures

The operative principle of a Fiber Optic Gyroscope (FOG) is based on the Sagnac effect. The
FOG is made of a fiber optic loop, fiber optic components, a photo-detector and a
semiconductor laser. The phase difference of the two light beams traveling in opposite
directions around the fiber optic loop is proportional to the rate of rotation of the fiber optic
loop. The rate information is internally integrated to provide the absolute measurements of
orientation. A FOG does not require frequent maintenance and have a longer lifetime of the
conventional mechanical gyroscopes. In a FOG the drift is also low. A complete analysis of
the accuracy and performances of this internal sensor has been developed in (Killian, 1994;
Borenstein & Feng, 1996; Zhu et al., 2000; Chung et al., 2001). This internal sensor represents
a simple low cost solution for producing accurate pose estimation of a mobile robot. The
FOG readings are denoted by y () = g ( ) + () , where g ( ) is the true value and ( ) is
an independent, zero mean, gaussian white sequence

2.3 Laser scanner measures

The distance readings by the Laser Measurement System (LMS) are related to the in-door
environment model and to the configuration of the mobile robot.
Denote with l the distance between the center of the laser scanner and the origin O of the
coordinate system ( O, X , Y ) fixed to the mobile robot, as reported in Fig. 2.

Fig. 2. Laser scanner displacement.

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

219

At the sampling time tk , the position xs , ys and orientation s of the center of the laser
scanner, referred to the inertial coordinate system (O , X , Y ) , have the following form:
xs ( tk ) = x ( tk ) + l cos ( tk )

(15)

y s ( tk ) = y ( tk ) + l sin ( tk )

(16)

s ( tk ) = ( t k )

(17)

The walls and the obstacles in an in-door environment are represented by a proper set of
planes orthogonal to the plane XY of the inertial coordinate system. Each plane P j ,
j {1, 2,K , n p } (where n p is the number of planes which describe the indoor environment),
is represented by the triplet Pr j , Pn j and P j , where Pr j is the normal distance of the plane
from the origin O , Pn j is the angle between the normal line to the plane and the X -direction
and P j is a binary variable, P j {1,1} , which defines the face of the plane reflecting the
laser beam. In such a notation, the expectation of the

i -th ( i = 1, 2,K, ns )

laser reading

j
d i j ( tk ) , relative to the present distance of the center of the laser scanner from the plane P ,

has the following expression (see Fig. 3):

d i j ( tk ) =

P j Pr j xs ( tk ) cos Pnj y s ( tk sin Pn j )

cos i

(18)

where

i j = Pn j i

(19)

with i [ 0 , 1 ] given by (see Fig. 4):

(20)
.
2
The vector composed of geometric parameters Pr j , Pn j and P j , j {1, 2,K , n p } , is denoted by .

i = s + i

Fig. 3. Laser scanner measure.

www.intechopen.com

220

Mobile Robots, Perception & Navigation

Fig. 4. Laser scanner field of view for plane P j .

The laser readings y j ( t ) are denoted by y j ( ) = d i j ( ) + i ( ) , where d i j () is the true value
k
d
d
i

expressed by (18) and i ( ) is an independent, zero mean, gaussian white sequence

.

3. Relative approaches for mobile robot localization

The purpose of this section is to propose and to compare three different algorithms for the mobile
robot localization only using internal sensors like odometers and gyroscopes. The first method
(Algorithm 1) is the simplest one and is merely based on a numerical integration of the raw
encoder data; the second method (Algorithm 2) replaces the gyroscopic data into the equations
providing the numerical integration of the increments provided by the encoders. The third
method (Algorithm 3) operates in a stochastic framework where the uncertainty originates by the
measurement noise and by the robot model inaccuracies. In this context the right approach is to
formulate the localization problem as a state estimation problem and the appropriate tool is the
EKF (see e.g. (Barshan & Durrant-Whyte, 1995; Garcia et al., 1995; Kobayashi et al., 1995; Jetto et al.,
1999; Sukkarieh et al., 1999; Roumeliotis & Bekey, 2000; Antoniali & Oriolo, 2001; Dissanayake et
al., 2001)). Hence, Algorithm 3 is a suitably defined EKF fusing together odometric and gyroscopic
data. In the developed solution, the dynamical equation of the state-space form of the robot
kinematic model, has been considered. The numerical integration equations of the encoder
increments have been considered for deriving the measure equation. This allows to fuse together
all the available informative content which is carried both by the robot dynamics and by the
acquired measures.
3.1 Algorithm 1
Equations (6)-(8) have been used to estimate the position and orientation of the mobile robot
at time tk +1 replacing the true values of ( tk ) tk and ( tk ) tk with their measures y ( tk )
and y ( tk ) respectively, provided by the encoders. An analysis of the accuracy of this

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

221

estimation procedure has been developed in (Wang, 1988; Martinelli, 2002), where it is
shown that the incremental errors on the encoder readings especially affect the estimate of
the orientation ( tk ) and reduce its applicability to short trajectories.
3.2 Algorithm 2
This algorithm is based on the ascertainment that the angular measure y ( tk ) provided by
the FOG is much more reliable than the orientation estimate obtainable with Algorithm 1.
Hence, at each time instant, Algorithm 2 provides an estimate of the robot position and
orientation x ( tk +1 ) , y ( tk +1 ) , y ( tk +1 ) , where y ( tk +1 ) is the FOG reading, x ( tk +1 ) and y ( tk +1 )

are computed through equations (6), (7), replacing ( tk ) tk with its measure y ( tk ) , ( tk )
with y ( tk ) and ( tk ) tk with y ( tk +1 ) y ( tk ) .
3.3 Algorithm 3
This algorithm operates in a stochastic framework exploiting the same measures of
Algorithm 2. A state-space approach is adopted with the purpose of defining a more
general method merging the information carried by the kinematic model with that
provided by the sensor equipment. The estimation algorithm is an EKF defined on the
basis of a state equation derived from (1)-(3) and of a measure equation inglobing the
incremental measures of the encoders y ( tk ) and the angular measure of the gyroscope
y ( tk ) . This is a difference with respect to other existing EKF based approaches,

(Barshan & Durrant-Whyte, 1995; Kobayashi et al., 1995; Sukkarieh et al., 1999;
Roumeliotis & Bekey, 2000; Antoniali & Oriolo, 2001; Dissanayake et al., 2001b), where
equations (1)-(3) are not exploited and the dynamical equation of the state-space model
is derived from the numerical integration of the encoder measures.
Denote with X ( t ) := x ( t ) , y ( t ) , ( t ) T the true robot state and with U ( t ) := ( t ) , ( t ) T the

robot control input. For future manipulations it is convenient to partition X ( t ) as

T
T
X ( t ) := X 1 ( t ) , ( t ) , with X 1 ( t ) := x ( t ) , y ( t ) . The kinematic model of the robot can be

written in the compact form of the following stochastic differential equation

dX ( t ) = F ( X ( t ) ,U ( t ) ) dt + d ( t )

(21)

where F ( X ( t ) ,U ( t ) ) represents the set of equations (1)-(3) and ( t ) is a Wiener process

such that E d ( t ) d ( t )T = Qdt . Its weak mean square derivative d ( t ) dt is a white noise
representing the model inaccuracies (parameter uncertainties, slippage,
process
dragging). It is assumed that Q = 2 I 3 , where I n denote the n n identity matrix. The
diagonal form of Q understands the hypothesis that model (21) describes the true dynamics
of the three state variables with nearly the same degree of approximation and with
independent errors.
Let tk = T be the constant sampling period and denote tk by kT , assume

www.intechopen.com

222

Mobile Robots, Perception & Navigation

U ( t ) = U ( kT ) := U ( k ) , for t kT , ( k + 1) T and denote by X ( k ) and by X ( k , k ) the

current state and its filtered estimate respectively at time instant tk = kT . Linearization
of (15) about U ( k 1) and X ( k , k ) and subsequent discretization with period T results
in the following equation
(22)
X ( k + 1) = Ad ( k ) X ( k ) + L ( k ) U ( k ) + D ( k ) + W ( k )
Partitioning vectors and matrices on the right hand side of equation (22) according to the
partition of the state vector one has

A1,1 ( k )
Ad ( k ) = exp ( A ( k ) T ) = d
A2,1d ( k )

A1,2d ( k )
L1 ( k )
D1 ( k )
, L (k ) =
, D (k ) =

A2,2d ( k )
L2 ( k )
D2 ( k )

0 0 ( k 1) sin ( k , k )

F ( X ( t ) ,U ( t ) )
A ( k ) :=
= 0 0 ( k 1) cos ( k , k )

X ( t )

X (t )= X ( k ,k ) 0 0

0
U ( t ) =U ( k 1)

(23)

(24)

( k 1) sin ( k , k ) T
1 0
:= I 2 , A1,2d ( k ) =
A1,1d ( k ) =

0 1
( k 1) cos ( k , k ) T

(25)

A2,1d ( k ) = [ 0 0] , A2,2d ( k ) = 1

(26)

T cos ( k , k ) 0.5 ( k 1) T 2 sin ( k , k )

L1 ( k ) =
, L2 ( k ) = [0 T ]
2
T sin ( k , k ) 0.5 ( k 1) T cos ( k , k )
T ( k 1) ( k , k ) sin ( k , k )
D1 ( k ) =
, D2 ( k ) = 0
T ( k 1) ( k , k ) cos ( k , k )

( k +1)T

W ( k ) :=

kT

W ( k )
exp A ( k ) ( k + 1) zT ( ) d = 1

W2 ( k )

(27)

(28)

(29)

with W1 ( k ) , W2 ( k ) , k = 0,1, 2,... .

2

The integral term W ( k ) given (29) has to be intended as a stochastic Wiener integral, its

covariance matrix is E W ( k ) W ( k )T := Q ( k ) = 2 ( k ) Q ( k ) , where

d

Q ( k ) Q1,2 ( k )
Q ( k ) = 1,1

Q2,1 ( k ) Q2,2 ( k )

T3
T3
T + 2 ( k 1) sin 2 ( k , k )
2 ( k 1) cos ( k , k ) sin ( k , k )

3
3
Q1,1 ( k ) =

T3
T3
2

2
2
( k , k ) sin ( k , k )
k
T
k
k
k
1
cos
1
cos
,

(
)
(
)
(
)

3
3

www.intechopen.com

(30)

(31)

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

223

T2

( k 1) 2 sin ( k , k )
T
Q1,2 ( k ) =
, Q2,1 ( k ) = Q1,2 ( k ) , Q2,2 ( k ) = T.
T2

( k 1) cos ( k , k )

(32)

Denote by Z ( k ) = z1 ( k ) , z2 ( k ) T the measurement vector at time instant kT , the elements of

Z ( k ) are: z1 ( k ) y ( tk ) , z2 ( k ) y ( tk ) , where y ( tk ) is the measure related to the

increments provided by the encoders through equations (9) and (13), y ( tk ) is the angular

measure provided by the FOG. The observation noise V ( k ) = ( k ) , ( k ) T is a white

where R = diag 2 , 2 . The diagonal form of R follows by the

sequence

independence of the encoder and FOG measures. As previously mentioned, the measure

z2 ( k ) provided by the FOG is much more reliable than z1 ( k ) , so that

. This gives

rise to a nearly singular filtering problem, where singularity of R arises due to the very
high accuracy of a measure. In this case a lower order non singular EKF can be derived
assuming that the original R is actually singular (Anderson & Moore, 1979). In the present
problem, assuming 2 = 0 , the nullity of R is m = 1 and the original singular EKF of order

n = 3 can be reduced to a non singular problem of order n m = 2 , considering the third

component ( k ) of the state vector X ( k ) coinciding with the known deterministic signal
z2 ( k ) = g ( k ) . Under this assumption, only X 1 ( k ) needs be estimated as a function of z1 ( ) .
As the measures z1 ( ) provided by the encoders are in terms of increments, it is convenient to

define the following extended state X ( k ) := X ( k )T , X ( k 1)T in order to define a measure

1
1

equation where the additive gaussian noise is white. The dynamic state-space equation for X ( k )
T

is directly derived from (22), taking into account that, by the assumption on z2 ( ) , in all vectors
and matrices defined in (25)(32), the term ( k , k ) must be replaced by ( k ) .
g

The following equation is obtained

X ( k + 1) = A ( k ) X ( k ) + L ( k ) U ( k ) + B ( k ) g ( k ) + D ( k ) + W ( k )
where

I2
A (k ) =
I2

A1,2d ( k )
02,2
L ( k )
, L (K ) = 1

, B (k ) =
02,2
02,1
02,2

D ( k )
W1 ( k )
D (k ) = 1
, W (k ) =

0
2,1
02,1

(33)

(34)
(35)

0i , j being the ( i j ) null matrix.

Equations (6), (7) and (13) and the way the state vector X ( k ) is defined imply that the

z1 ( k ) y ( tk ) can be indifferently expressed as

www.intechopen.com

224

or

Mobile Robots, Perception & Navigation

1
1
z1 ( k ) = ( k ) ,0, ( k ) ,0 X ( k ) + ( k )

(36)

1
1
z1 ( k ) = 0, ( k ) ,0, ( k ) X ( k ) + ( k )

(37)

where

( kT ) :=

sin

( tk ) tk
2

( tk ) tk
2

(38)

(39)
( tk ) tk

2
sin ( tk ) +

( tk ) tk
2

2
and ( tk ) = g ( tk ) . The measure equations (36) and (37) can

( kT ) :=
with ( tk ) tk = g ( tk +1 ) g ( tk )

( tk ) tk

cos ( tk ) +

sin

( tk ) tk

be combined to obtain a unique equation where z1 ( k ) is expressed as a function both of

x ( k + 1) x ( k ) and of y ( k + 1) y ( k ) . As the amount of observation noise is the same,

equations (36) and (37) are averaged, obtaining

z1 ( k ) = C1 ( k ) X ( k ) + v1 ( k )

(40)

where C ( k ) = ( k ) 1 2, ( k ) 1 2, ( k ) 1 2, ( k ) 1 2 and v1 ( k ) := ( k ) . Equations (33)

1

and (40) represent the linearized, discretized state-space form to which the classical EKF
algorithm has been applied.
3.4 Experimental results
The experimental tests have been performed on the TGR Explorer powered wheelchair
(TGR Bologna, 2000) in an indoor environment. This mobile base has two driving wheels
and a steering wheel. The odometric system is composed by two optical encoders connected
to independent passive wheels aligned with the axes of the driving wheels, as shown in Fig.
5. A sampling time of 0.4 s has been used.

Fig. 5. TGR Explorer with data acquisition system for FOG sensor and incremental encoders.

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

225

The odometric data are the incremental measures that at each sampling interval are
provided by the encoders attached to the right and left passive wheels. The incremental
optical encoders SICOD mod. F3-1200-824-BZ-K-CV-01 have been used to collect the
odometric data. Each encoder has 1200 pulses/rev. and a resolution of 0.0013 rad. These
measures are directly acquired by the low level controller of the mobile base. The gyroscopic
measures on the absolute orientation have been acquired in a digital form by a serial port on
the on-board computer. The fiber optic gyroscope HITACHI mod. HOFG-1 was used for
measuring the angle of the mobile robot. The main characteristics of this FOG are
reported in the Table 1. While the used FOG measures the rotational rates with a very high
accuracy, the internal integration of angular rates to derive the heading angle can suffer
from drift (Barshan & Durrant-Whyte, 1995; Komoriya & Oyama, 1994). Because of the low
rate integration drift of the used FOG (see Table 1), the drift is not accounted for in the
proposed experiments where the robot task duration is on the order of several minutes. For
longer task duration the rate integration drift can be compensated as proposed in (Ojeda et
al., 2000) or can be periodically reset by a proper docking system or an absolute sensing
mechanism (Barshan & Durrant-Whyte, 1995).
Rotation Rate
Angle Measurement Range
Random Walk

-1.0472 to +1.0472 rad/s

-6.2832 to +6.2832 rad
0.0018 rad h

Zero Drift (Rate Integration)

0.00175rad h

Non-linearity of Scale Factor

within 1.0%
Time Constant
Typ. 20 ms
Response Time
Typ. 20 ms
Data Output Interval
Min. 10 ms
Warm-up Time
Typ. 6.0 s
Table 1. Characteristics of the HITACHI gyroscope mod. HFOG - 1.
The navigation module developed for the considered mobile base interacts with the user in
order to involve her/him in the guidance of the vehicle without limiting the functionality
and the security of the system. The user sends commands to the navigation module through
the user interface and the module translates the user commands in the low level command
for the driving wheels. Two autonomy levels are developed to perform a simple filtering or
to introduce some local corrections of the user commands on the basis of the environment
information acquired by a set of sonar sensors (for more details see (Fioretti et al., 2000)).
The navigation system is connected directly with the low level controller and with the Fiber
Optic Gyroscope by analog and digital converters and serial port RS232, respectively.
All the experiments have been performed making the mobile base track relatively long trajectories.
In the indoor environment of our Department a closed trajectory of 108 m length, characterized by
a lot of orientation changes has been considered. The trajectory has been imposed by the user
interface with the end configuration coincident with the start configuration. In order to quantify
the accuracy of the proposed localization algorithms, six markers have been introduced along the
trajectory. The covariance matrix R of the observation noise V () has been determined by an
analysis of the sensor characteristics. The detected estimate errors in correspondence of the marker

www.intechopen.com

226

Mobile Robots, Perception & Navigation

configurations (the distance between the marker and the corresponding estimated configuration)
of the mobile base with Algorithm 1 have been reported in the first row of Table 2. This algorithm
fails to successfully localize the robot, because as it was predictable, the results exhibit a very large
drift and the estimated trajectory is totally wrong after few meters of travel.
With reference to the same experimental path, the trajectory estimated by Algorithm 2 is
more accurate with respect to that estimated by Algorithm 1. Algorithm 2 successfully
removes the integration error present in the odometry. The goodness of the estimated
trajectory is quantified by the numerical values of the estimation errors in correspondence of
the markers. These values are reported in the second row of Table 2.
The experimental results obtained by Algorithm 3 are relatively close to those of Algorithm
2. The improvement introduced by Algorithm 3 can be evaluated looking at the numerical
values reported in the third row of Table 2.
Markers
Mk1
Mk2
Mk3
Mk4
Mk5
Mk6
stop
Algorithm 1
0.014
0.143
0.690
4.760
1.868
3.770
6.572
Algorithm 2
0.012
0.041
0.042
0.164
0.142
0.049
0.187
Algorithm 3
0.012
0.037
0.035
0.150
0.106
0.030
0.161
Table 2. Estimation errors (in meters) in correspondence of the marker configurations
(distance between the marker and the corresponding estimated configuration).
3.5 Comments
The performed experimental tests show that the simple odometric localization is not
satisfactory, making it necessary the introduction of another internal sensor. A fiber optic
gyroscope showed to be a key tool for obtaining a significant improvement in the accuracy
of the estimated trajectory. Algorithm 2 is very similar to Algorithm 1, the only difference is
that Algorithm 2 exploits the gyroscopic measures. This is enough to produce a huge
improvement of the estimated trajectory, thus confirming the validity of Equations (6), (7)
provided that an accurate estimate of the robot orientation is available.
Algorithm 3 uses the same measures of Algorithm 2 but operates in the stochastic framework of
the Kalman filtering theory. The novelty of the proposed EKF is that its formulation explicitly
includes both the information carried by the model of the robot and the information carried by
the observations. This introduces a further improvement with respect to Algorithm 2 and a very
high degree of accuracy in the estimated trajectory is achieved. The main merit of Algorithm 3 is
that it operates in a state-space form where sensor and model uncertainties are intrinsically taken
into account. This makes the estimator more robust with respect to possible uncertain physical
parameters and/or not exactly known initial conditions. Taking also into account its modest
computational burden, Algorithm 3 appears to be the most appealing among the three
localization procedures here proposed.

4. Absolute approaches for mobile robot localization

The purpose of this section is to propose and to experimentally evaluate a localization algorithm
based on a measure apparatus composed of a set of internal and external sensors of a different
nature and characterized by a highly different degree of accuracy. The sensor equipment
includes odometric, gyroscopic and laser measures.
The main technical novelty of this section is the integration in a stochastic framework of

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

227

the new set of measures. Both the information carried by the kinematic model of the
robot and that carried by the dynamic equations of the odometry are exploited. The
nearly singular filtering problem arising from the very high accuracy of angular
measure has been explicitly taken into account. An exteroceptive laser sensor is
integrated for reducing the continuous growth in the integrated error affecting any
relative localization algorithm, such as the Algorithm 3.
4.1 Algorithm 4
The algorithm operates in a stochastic framework as Algorithm 3, and is based on the
ascertainment that the angular measure y ( tk ) provided by the FOG is much accurate than
the other measures. This gives rise to a nearly singular filtering problem which can be
solved by a lower order non singular Extended Kalman Filter, as described in subsection 3.3.
The EKF is defined on the basis of a state equation derived from (1)(3) and of a measure
equation containing the incremental measures of the encoders y ( tk ) and the distance
measures y j ( t ) , i = 1, 2,K , ns , provided by the laser scanner from the P j plane,
k
d
i

j {1, 2,K , n p } . The angular measure of the gyroscope y ( tk ) is assumed coincident to the
third component ( k ) of the state vector X ( k ) .
Let Z ( k ) be the measurement vector at time instant kT . Its dimension is not constant,
depending on the number of sensory measures that are actually used at each time instant.
The measure vector Z ( k ) is composed by two subvectors Z1 ( k ) = z1 ( k ) , z2 ( k ) T and

Z 2 ( k ) = z3 ( k ) , z4 ( k ) ,L , z2+ ns ( k ) , where the elements of

T

Z1 ( k )

are:

z1 ( k ) y ( k ) ,

z2 ( k ) y ( k ) , where y ( k ) is the measure related to the increments provided by the

encoders through equations (9) and (13), y ( k ) is the angular measure provided by the
FOG. The elements of Z 2 ( k ) are: z2+i ( k ) = d i j ( k ) + i ( k ) , i = 1, 2,K , ns , j {1, 2,K , n p } , with

di j ( k ) given by (18) and

. The environment map provides the information needed

j

to detect which is the plane P in front of the laser.

T
The observation noise V ( k ) = ( k ) , ( k ) ,1 ( k ) ,K ,n ( k ) , is a white sequence
s

where R := block diag [ R1 , R2 ] , with R1 := diag 2 , 2 and R2 := diag 12 , 22 ,L , n2 .

The diagonal form of R follows by the independence of the encoder, FOG and laser
scanner measures.
The components of the extended state vector X ( k ) and the last

ns

components of vector

Z ( k ) are related by a non linear measure equation which depends on the environment

geometric parameter vector . The dimension ns ( k ) is not constant, depending on the

number of laser scanner measures that are actually used at each time, this number depends
on enviroment and robot configuration.
Linearization of the measure equation relating Z 2 ( k ) and X ( k ) about the current estimate

www.intechopen.com

228

Mobile Robots, Perception & Navigation

of X ( k ) results in:

Z 2 ( k ) = C2 ( k ) X ( k ) + V2 ( k )

(41)

where V2 ( k ) = 1 ( k ) ,2 ( k ) ,K ,n ( k ) is a white noise sequence

s

T
T
T
C2 ( k ) := c1 ( k ) , c2 ( k ) ,L , cns ( k ) ( k )

with

ci ( k ) =

and

P j
cos Pn j , sin Pn j ,0,0 , i = 1, 2,K , ns ( k ) , j {1, 2,K , n p }
cos i j

(42)

(43)

and

i j = Pn j g i +

(44)

Equations (33), (40) and (41) represent the linearized, discretized state-space form to which
the classical EKF algorithm has been applied.
4.2 Laser scanner readings selection
To reduce the probability of an inadequate interpretation of erroneous sensor data, a
method is introduced to deal with the undesired interferences produced by the presence of
unknown obstacles on the environment or by incertitude on the sensor readings. Notice that
for the problem handled here both the above events are equally distributed. A simple and
efficient way to perform this preliminary measure selection is to compare the actual sensor
readings with their expected values. Measures are discharged if the difference exceeds a
time-varying threshold. This is here done in the following way: at each step, for each
measure z2+i ( k ) of the laser scanner, the residual i ( k ) = z2+i ( k ) d i j ( k ) represents the
difference between the actual sensor measure z2+i ( k ) and its expected value d i j ,

i = 1, 2,K, ns ( k ) , j = 1, 2,K , n p , which is computed by (18) on the basis of the current estimate
of the vector state X ( k ) . As

, the current value z2+i ( k ) is accepted if

i ( k ) 2 si ( k ) (Jetto et al., 1999). Namely, the variable threshold is chosen as two times the
standard deviation of the innovation process.
4.3 Experimental results
The experimental tests have been performed in an indoor environment using the same TGR
Explorer powered wheelchair (TGR Bologna, 2000), described in Section 3.4.
The laser scanner measures have been acquired by the SICK LMS mod. 200 installed on the
vehicle. The main characteristics of the LMS are reported in Table 3.
Aperture Angle
Angular Resolution
Response Time
Resolution

www.intechopen.com

3.14 rad
0.0175/ 0.0088/ 0.0044 rad
0.013/ 0.026/ 0.053 s
0.010 m

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

Systematic Error
Statistic Error (1 Sigma)
Laser Class
Max. Distance
Transfer Rate
Table 3. Laser.

229

0.015 m
0.005 m
1
80 m
9.6/ 19.2/ 38.4/ 500 kBaud

A characterization study of the Sick LMS 200 laser scanner has been performed as proposed
in (Ye & Borenstein, 2002). Different experiments have been carried out to analyze the effects
of data transfer rate, drift, optical properties of the target surfaces and incidence angle of the
laser beam. Based on empirical data a mathematical model of the scanner errors has been
obtained. This model has been used as a calibration function to reduce measurement errors.
The TGR Explorer powered wheelchair with data acquisition system for FOG sensor,
incremental encoders, sonar sensors and laser scanner is shown in Fig. 5.

Fig. 6. Sample of the estimated trajectory. The dots are the actually used laser scanner measures.

www.intechopen.com

230

Mobile Robots, Perception & Navigation

Alg 4

Alg 3

A significative reduction of the wrong readings produced by the presence of unknown

obstacles has been realized by the selection of the laser scanner measures using the
procedure described in the previous subsection .
Different experiments have been performed making the mobile base track short and
relatively long and closed trajectories. Fig. 6 illustrates a sample of the obtained results; the
dots in the figure, are the actually used laser scanner measures. In the indoor environment
of our Department, represented by a suitable set of planes orthogonal to the plane XY of
the inertial system, a trajectory of 118 m length, characterized by orientation changes, has
been imposed by the user interface. The starting and final positions have been measured,
while six markers specify different middle positions; this permits to compute the distance
and angle errors between the marker and the corresponding estimated configuration.
In these tests, the performances of Algorithm 4 have been compared with those ones of the
Algorithm 3, which is the most appealing among the three relative procedures here
analyzed. Table 4 summarizes the distance and angle errors between the marker and the
corresponding configurations estimated by the two algorithms.
Mk1
0.1392

Mk2
0.095

Mk3
0.2553

0.49

0.11

0.85

Error

0.0156

0.0899

0.59

0.05

Error

Markers
Mk4
0.1226

Mk5
0.2004

Mk6
0.0301

0.58

1.39

0.84

0.0659

0.1788

0.0261

0.0601

0.45

0.07

0.72

0.12

stop
0.3595
2.66
0.0951
1.55

Table 4. Estimation errors (in meters) in correspondence of the marker configurations

(distance between the marker and the corresponding estimated configuration) and
corresponding angular errors (in degrees).
Other significant sets of experiments have been performed inside a room, considering a
short trajectory of 20 m characterized by different orientation changes (see Fig. 7).

Fig. 7. Sample of the estimated trajectory inside the room, where dots indicate the laser measures.

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

231

Alg 4

The room has been modelled very carefully, permitting a precise evaluation of the distance
and angle errors between the final position and the corresponding configuration estimated
by the Algorithm 4; Table 5 resumes these results.
error

final position
0.0061

0.27

Table 5. Estimation distance errors (in meters) and corresponding angular errors (in degrees).
In order to investigate further the efficiency of the developed Algorithm 4 and to evaluate its
correction performances, it has been imposed a wrong initial position (see Table 6 and Fig. 8).
error of initial position error of final position
error

0.2236
1.5

0.0152
0.73

Table 6. Distance (in meters) and angle (in degrees) errors introduced on the initial position
and corresponding errors on the final position.

Fig. 8. Estimated trajectory with a wrong initial positioning.

As a result, it has been seen that the Algorithm 4 is able to correct possible errors on the
initial positioning, as confirmed by the results reported in Table 6.
4.4 Comments
As shown by the developed experimental tests (see Table 4), Algorithm 4 permits to obtain a
much more reliable and accurate positioning than that one obtained by Algorithm 3. Note
that estimation errors on the final position of the Algorithm 3 are due to the angle drift
introduced by the gyroscope.
Additionally, Algorithm 4 improves the positioning accuracy in spite of a wrong initial
positioning. Table 6 shows as the possible errors introduced by a wrong initial pose, have
been efficiently corrected by the Extended Kalman Filter.

www.intechopen.com

232

Mobile Robots, Perception & Navigation

5. Concluding remarks
This chapter has presented a concise look at the problems and methods relative to the
mobile robot localization. Both the relative and absolute approaches have been discussed.
Relative localization has the main advantage of using a sensor equipment which is totally
self-contained in the robot. It is relatively simple to be used and guarantees a high data rate.
The main drawback is that the localization errors may considerably grow over time.
The three corresponding algorithms which have been proposed only use odometric and
gyroscopic measures. The experimental tests relative to Algorithm 1 show that the
incremental errors of the encoder readings heavily affect the orientation estimate, thus
reducing the applicability of the algorithm to short trajectories. A significant improvement is
introduced by Algorithm 2 where the odometric measures are integrated with the angular
measures provided by a gyroscope.
Algorithm 3 uses the same measures of Algorithm 2 but operates in a stochastic framework.
The localization problem is formulated as a state estimation problem and a very accurate
estimate of the robot localization is obtained through a suitably defined EKF. A further
notable improvement is provided by the fusion of the internal measures with absolute laser
measures. This is clearly evidenced by Algorithm 4 where an EKF is again used.
A novelty of the EKF algorithms used here is that the relative state-space forms include all
the available information, namely both the information carried by the vehicle dynamics and
by the sensor readings. The appealing features of this approach are:

The possibility of collecting all the available information and uncertainties of a

different kind in the compact form of a meaningful state-space representation,

The recursive structure of the solution,

The modest computational effort.

Other previous, significant experimental tests have been performed at our Department using
sonar measures instead of laser readings (Bonci et al., 2004; Ippoliti et al., 2004). Table 7 reports a
comparison of the results obtained with Algorithm 3, Algorithm 4, and the algorithm (Algorithm
4(S)) based on an EKF fusing together odometric, gyroscopic and sonar measures. The
comparative evaluation refers to the same relatively long trajectory used for Algorithm 4.
Alg 3
Alg 4
Alg 4(S)
0.8079
0.0971
0.1408
2. 4637
0.7449
1. 4324

Table 7. Estimation errors (in meters) in correspondence of the final vehicle configuration
(distance between the actual and the corresponding estimated configuration) and
corresponding angular errors (in degrees).
error

Table 7 evidences that in spite of a higher cost with respect to the sonar system, the
localization procedure based on odometric, inertial and laser measures does really seem to
be an effective tool to deal with the mobile robot localization problem.
A very interesting and still open research field is the Simultaneous Localization and Map
Building (SLAM) problem. It consists in defining a map of the unknown environment and
simultaneously using this map to estimate the absolute location of the vehicle. An efficient
solution of this problem appears to be of a dominant importance because it would definitely
confer autonomy to the vehicle. The SLAM problem has been deeply investigated in
(Leonard et al., 1990; Levitt & Lawton, 1990; Cox, 1991; Barshan & Durrant-Whyte, 1995;
Kobayashi et al., 1995; Thrun et al., 1998; Sukkarieh et al., 1999; Roumeliotis & Bekey, 2000;

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

233

Antoniali & Orialo, 2001; Castellanos et al., 2001; Dissanayake et al., 2001a; Dissanayake et
al., 2001b; Zunino & Christensen, 2001; Guivant et al., 2002; Williams et al., 2002; Zalama et
al., 2002; Rekleitis et al., 2003)). The algorithms described in this chapter, represent a solid
basis of theoretical background and practical experience from which the numerous
questions raised by SLAM problem can be solved, as confirmed by the preliminary results in
(Ippoliti et al., 2004; Ippoliti et al., 2005).

6. References
Anderson, B.D.O. & Moore, J.B. (1979). Optimal Filtering. Prentice-Hall, Inc, Englewood
Cliffs
Antoniali, F.M. & Oriolo, G. (2001). Robot localization in nonsmooth environments:
experiments with a new filtering technique, Proceedings of the IEEE International
Conference on Robotics and Automation (2001 ICRA), Vol. 2, pp. 15911596
Arras, K.O.; Tomatis, N. & Siegwart, R. (2000). Multisensor on-the-fly localization using
laser and vision, Proceedings of the 2000 IEEE/RSJ International Conference
onIntelligent Robots and Systems, (IROS 2000), Vol. 1, pp. 462467
Barshan, B. & Durrant-Whyte, H.F. (1995). Inertial navigation systems for mobile robots.
IEEE Transactions on Robotics and Automation, Vol. 11, No. 3, pp. 328342
Bemporad, A.; Di Marco, M. & Tesi, A. (2000). Sonar-based wall-following control of
mobile robots. Journal of dynamic systems, measurement, and control, Vol. 122, pp.
226230
Bonci, A.; Ippoliti, G.; Jetto, L.; Leo, T. & Longhi, S. (2004). Methods and algorithms for
sensor data fusion aimed at improving the autonomy of a mobile robot. In:
Advances in Control of Articulated and Mobile Robots, B. Siciliano, A. De Luca ,
C. Melchiorri, and G. Casalino, Eds. Berlin, Heidelberg, Germany: STAR (Springer
Tracts in Advanced Robotics ), Springer-Verlag, Vol. 10, pp. 191222.
Borenstein, J. & Feng, L. (1996). Measurement and correction of systematic odometry errors
in mobile robots. IEEE Transaction on Robotics and Automation, Vol. 12, No. 6, pp.
869880
Borenstein, J.; Everett, H. R.; Feng, L. & Wehe, D. (1997). Mobile robot positioning: Sensors
and techniques. Journal of Robotic Systems, Vol. 14, No. 4, pp. 231249
Castellanos, J.A.; Neira, J. & Tards, J.D. (2001). Multisensor fusion for simultaneous
localization and map building. IEEE Transactions on Robotics and Automation, Vol.
17, No. 6, pp. 908914
Chung, H.; Ojeda, L. & Borenstein, J. (2001). Accurate mobile robot dead-reckoning with a
precision-calibrated fiber-optic gyroscope. IEEE Transactions on Robotics and
Automation, Vol. 17, No. 1, pp. 8084
Cox, I. (1991). Blanche an experiment in guidance and navigation of an autonomous
robot vehicle. IEEE Transactions on Robotics and Automation, Vol. 7, No. 2, pp. 193
204
De Cecco, M. (2003). Sensor fusion of inertial-odometric navigation as a function of the
actual manoeuvres of autonomous guided vehicles. Measurement Science and
Tecnology, Vol. 14, pp. 643653
Dissanayake, M.; Newman, P.; Clark, S.; Durrant-Whyte, H.F. & Csorba, M. (2001a). A
solution to the simultaneous localization and map building (slam) problem. IEEE
Transactions on Robotics and Automation, Vol. 17, No. 3, pp. 229241

www.intechopen.com

234

Mobile Robots, Perception & Navigation

Dissanayake, G.; Sukkarieh, S.; Nebot, E. & Durrant-Whyte, H.F. (2001b). The aiding of a low-cost
strapdown inertial measurement unit using vehicle model constraints for land vehicle
applications. IEEE Transactions on Robotics and Automation, Vol. 17, pp. 731747
Durrant-Whyte, H.F. (1988). Sensor models and multisensor integration. International Journal
of Robotics Research, Vol. 7, No. 9, pp. 97113
Fioretti, S.; Leo, T. & Longhi, S. (2000). A navigation system for increasing the autonomy and
the security of powered wheelchairs. IEEE Transactions on Rehabilitation Engineering,
Vol. 8, No. 4, pp. 490498
Garcia, G.; Bonnifait, Ph. & Le Corre, J.-F. (1995). A multisensor fusion localization
algorithm with self-calibration of error-corrupted mobile robot parameters,
Proceedings of the International Conference in Advanced Robotics, ICAR95, pp. 391397,
Barcelona, Spain
Gu, J.; Meng, M.; Cook, A. & Liu, P.X. (2002). Sensor fusion in mobile robot: some
perspectives, Proceedings of the 4th World Congress on Intelligent Control and
Automation, Vol. 2, pp. 11941199
Guivant, J.E.; Masson, F.R. & Nebot, E.M. (2002). Simultaneous localization and map
building using natural features and absolute information. In: Robotics and
Autonomous Systems, pp. 7990
Ippoliti, G.; Jetto, L.; La Manna, A. & Longhi, S. (2004). Consistent on line estimation of
environment features aimed at enhancing the efficiency of the localization
procedure for a powered wheelchair, Proceedings of the Tenth International
Symposium on Robotics with Applications - World Automation Congress (ISORA-WAC
2004), Seville, Spain
Ippoliti, G.; Jetto, L.; La Manna, A. & Longhi, S. (2005). Improving the robustness properties
of robot localization procedures with respect to environment features uncertainties,
Proceedings of the IEEE International Conference on Robotics and Automation (ICRA
2005), Barcelona, Spain
Jarvis, R.A. (1992). Autonomous robot localisation by sensor fusion, Proceedings of the IEEE
International Workshop on Emerging Technologies and Factory Automation, pp. 446450
Jetto, L.; Longhi, S. & Venturini, G. (1999). Development and experimental validation of an
adaptive extended Kalman filter for the localization of mobile robots. IEEE
Transactions on Robotics and Automation, Vol. 15, pp. 219229
Killian, K. (1994). Pointing grade fiber optic gyroscope, Proceedings of the IEEE Symposium on
Position Location and Navigation, pp. 165169, Las Vegas, NV,USA
Kobayashi, K.; Cheok, K.C.; Watanabe, K. & Munekata, F. (1995). Accurate global
positioning via fuzzy logic Kalman
filter-based sensor fusion technique,
Proceedings of the 1995 IEEE IECON 21st International Conference on Industrial
Electronics, Control, and Instrumentation, Vol. 2, pp. 11361141, Orlando, FL ,USA
Komoriya, K. & Oyama, E. (1994). Position estimation of a mobile robot using optical fiber
gyroscope (ofg), Proceedings of the 1994 International Conference on Intelligent Robots
and Systems (IROS94), pp. 143149, Munich, Germany
Leonard, J.; Durrant-Whyte, H.F. & Cox, I. (1990). Dynamic map building for autonomous
mobile robot, Proceedings of the IEEE International Workshop on Intelligent Robots and
Systems (IROS 90), Vol. 1, pp. 8996, Ibaraki, Japan

www.intechopen.com

Comparative Analysis of Mobile Robot Localization Methods Based On

Proprioceptive and Exteroceptive Sensors

235

Levitt, T.S. & Lawton, D.T. (1990). Qualitative navigation for mobile robots. Artificial
Intelligence Journal, Vol. 44, No. 3, pp. 305360
Mar, J. & Leu, J.-H. (1996). Simulations of the positioning accuracy of integrated vehicular
navigation systems. IEE Proceedings - Radar, Sonar and Navigation, Vol. 2, No. 143,
pp. 121128
Martinelli, A. (2002). The odometry error of a mobile robot with a synchronous drive
system. IEEE Transactions on Robotics and Automation, Vol. 18, No. 3, pp. 399405
Ojeda, L.; Chung, H. & Borenstein, J. (2000). Precision-calibration of fiber-optics gyroscopes
for mobile robot navigation, Proceedings of the 2000 IEEE International Conference on
Robotics and Automation, pp. 20642069, San Francisco, CA, USA
Panzieri, S.; Pascucci, F. & Ulivi, G. (2002). An outdoor navigation system using GPS
and inertial platform. IEEE/ASME Transactions on Mechatronics, Vol. 7, No. 2,
pp. 134142
Rekleitis, I.; Dudek, G. & Milios, E. (2003). Probabilistic cooperative localization and
mapping in practice, Proceedings of the IEEE International Conference on Robotics and
Automation (ICRA 03), Vol. 2, pp. 1907 1912
Roumeliotis, S.I. & Bekey, G.A. (2000). Bayesian estimation and Kalman filtering: a unified
framework for mobile robot localization, Proceedings of the IEEE International
Conference on Robotics and Automation (ICRA 00), Vol. 3, pp. 29852992, San
Francisco, CA, USA
Sukkarieh, S.; Nebot, E.M. & Durrant-Whyte, H.F. (1999). A high integrity IMU GPS
navigation loop for autonomous land vehicle applications. IEEE Transactions on
Robotics and Automation, Vol. 15, pp. 572578
Talluri, R. & Aggarwal, J.K. (1992). Position estimation for an autonomous mobile robot in
an outdoor environment. IEEE Transactions on Robotics and Automation, Vol. 8, No. 5,
pp. 573584
TGR Bologna (2000). TGR Explorer. Italy [Online]. Available on-line: http://www.tgr.it.
Thrun, S.; Fox, D. & Burgard, W. (1998). A probabilistic approach to concurrent mapping
and localization for mobile robots. Machine Learning Autonom. Robots, Vol. 31, pp.
2953. Kluwer Academic Publisher, Boston
Wang, C.M. (1988). Localization estimation and uncertainty analysis for mobile robots,
Proceedings of the Int. Conf. on Robotics and Automation, pp. 12301235
Williams, S.B.; Dissanayake, G. & Durrant-Whyte, H.F. (2002). An efficient approach to
the simultaneous localization and mapping problem, Proceedings of the 2002
IEEE International Conference on Robotics and Automation, pp. 406411,
Washington DC, USA
Ye, C. & Borenstein, J. (2002). Characterization of a 2d laser scanner for mobile robot obstacle
negotiation, Proceedings of the IEEE International Conference on Robotics and
Automation (ICRA 02), Vol. 3, pp. 2512 2518, Washington DC, USA
Yi, Z.; Khing, H. Y.; Seng, C.C. & Wei, Z.X. (2000). Multi-ultrasonic sensor fusion for mobile
robots, Proceedings of the IEEE Intelligent Vehicles Symposium, Vol. 1, pp. 387391,
Dearborn, MI, USA
Zalama, E.; Candela, G.; Gmez, J. & Thrun, S. (2002). Concurrent mapping and
localization for mobile robots with segmented local maps, Proceedings of the 2002
IEEE/RSJ Conference on Intelligent Robots and Systems, pp. 546551, Lausanne,
Switzerland

www.intechopen.com

236

Mobile Robots, Perception & Navigation

Zhu, R.; Zhang, Y. & Bao, Q. (2000). A novel intelligent strategy for improving measurement
precision of FOG. IEEE Transactions on Instrumentation and Measurement, Vol. 49,
No. 6, pp. 11831188
Zhuang, W. & Tranquilla, J. (1995). Modeling and analysis for the GPS pseudo-range
observable. IEEE Transactions on Aerospace and Electronic Systems, Vol. 31, No. 2, pp.
739751
Zunino, G. & Christensen, H.I. (2001). Simultaneous localization and mapping in domestic
environments, Proceedings of the International Conference on Multisensor Fusion and
Integration for Intelligent Systems (MFI 2001), pp. 6772

www.intechopen.com

Mobile Robots: Perception & Navigation

Edited by Sascha Kolski

ISBN 3-86611-283-1

Hard cover, 704 pages

Publisher Pro Literatur Verlag, Germany / ARS, Austria

Published online 01, February, 2007

Published in print edition February, 2007

Today robots navigate autonomously in office environments as well as outdoors. They show their ability to
beside mechanical and electronic barriers in building mobile platforms, perceiving the environment and
deciding on how to act in a given situation are crucial problems. In this book we focused on these two areas of
mobile robotics, Perception and Navigation. This book gives a wide overview over different navigation
techniques describing both navigation techniques dealing with local and control aspects of navigation as well
es those handling global navigation aspects of a single robot and even for a group of robots.

How to reference

In order to correctly reference this scholarly work, feel free to copy and paste the following:
Gianluca Ippoliti, Leopoldo Jetto, Sauro Longhi and Andrea Monteriu (2007). Comparative Analysis of Mobile
Robot Localization Methods Based on Proprioceptive and Exteroceptive Sensors, Mobile Robots: Perception &
Navigation, Sascha Kolski (Ed.), ISBN: 3-86611-283-1, InTech, Available from:
http://www.intechopen.com/books/mobile_robots_perception_navigation/comparative_analysis_of_mobile_rob
ot_localization_methods_based_on_proprioceptive_and_exteroceptive_

InTech Europe

University Campus STeP Ri

Slavka Krautzeka 83/A
51000 Rijeka, Croatia
Phone: +385 (51) 770 447
Fax: +385 (51) 686 166
www.intechopen.com

InTech China

Unit 405, Office Block, Hotel Equatorial Shanghai

No.65, Yan An Road (West), Shanghai, 200040, China
Phone: +86-21-62489820
Fax: +86-21-62489821