InternationalJournalofAdvancedRoboticSystems,Vol.7,No.
4(2010) 24
ISSN17298806,pp.2532
Research on Centroid Position
for Stairs Climbing Stability of
Search and Rescue Robot
Yan Guo
1,2
, Aiguo Song
1
, Jiatong Bao
1
, Huatao Zhang
1
, and Hongru Tang
3
1
SchoolofInstrumentScienceandEngineering,SoutheastUniversity,Nanjing210096,China
2
BeckmanInstitute,UniversityofIllinoisatUrbanaChampaign,Urbana61801,Illinois,USA
3
SchoolofEnergyandPowerEngineering,YangzhouUniversity,Yangzhou225009,China
CorrespondingauthorEmail:a.g.song@seu.edu.cnandy.guo@illinois.edu
Abstract:Thispaperrepresentstherelationshipbetweenthestabilityofstairsclimbingandthecentroidposition
of the search and rescue robot. The robot system is considered as a mass point plane model and the kinematics
featuresareanalyzedtofindtherelationshipbetweencentroidpositionandthemaximalpitchangleofstairsthe
robot could climb up. A computable function about this relationship is given in this paper. During the stairs
climbing,thereisamaximalstabilitykeepingangledependsonthecentroidpositionandthepitchangleofstairs,
and the numerical formula is developed about the relationship between the maximal stabilitykeeping angle and
thecentroidpositionandpitchangleofstairs.Theexperimentdemonstratesthetrustworthyandcorrectionofthe
methodinthepaper.
Keywords:centroid,stairsclimbing,stability,searchandrescuerobot
1.Introduction
Search and rescue robot is a special type robot which is
used in the search and rescue works after the nature or
manmade disasters, such as an earthquake, hurricane,
debris and mine collapse. Many researchers have paid
more attention on the search and rescue robots. Most
work is about the architecture of the robotic system and
some types of autonomous methods. The design
methodandarchitectureofasearchandrescuerobotare
given in recent years (Ko, A., Lau, H., 2009). Some
researchers focus on the automation navigation in the
fieldenvironment(Guo,Y.,Song,A.,Bao,J.,TangH.,Cui,
J.,2009).Identifiedfromtheothercommonmobilerobot,
the search and rescue robot always works under the
unstructured environment that means the complex and
changeable work situations, so the excellent traveling
ability is requested and domain its applications. In all
hard work situations which are research hot points and
difficult points, the stairs climbing is one of them. Relate
works always focus on how to climb the stairs and the
stability of the climbing process and the stability of the
robot during the climbing basically determines the
performanceofthesearchandrescuerobot.Somegroups
work for design a novel structure of the robot to achieve
thestairsclimbing(Woo,C.K.,Choi,H.D.,Yoon,S.,Kim,
S. H., Yoon, K. K., 2007), but the complex mechanisms
will lead to the reduction of reliability. Some groups are
workingtofindthemethodsthatautonomouslyclimbing
the stairs for the robot (Pinhas, B. T., Ito, S., Goldenberg,
A., 2009) (Kim, B. S., Vu, Q. S., Song, J. B., Yim, C. H.,
2010).However,theautonomousmethodshavetoomuch
limitation of the environment constrains, and it is weak
about their generalization ability in the real application.
Somegroupsanalyzethestairsclimbingabilitybasedon
the geometry and kinetic features of the robot (Liu, J.,
Wang, Y., Ma, S., Li, B., 2005) (Liu, Y., Liu, G., 2009).
Hence, some outside geometry and kinetic features are
not the only constrains that influence the stairs climbing
ability, in most situations, the position of centroid play
much more decisive role in the performance rather than
the others. The group of Kim analyzed the walking
balanceofalegrobotbasedoncentroid(Kim,B.H.,2009).
Because of the importance of centroid position, the
groups of Zhu and Wang research on the centroid
adjustmentmethodstoincreasethetravelingability(Zhu,
X. L., Wang, H. G., Fang, L. J., Zhao, M. Y., Wang, L.D.,
2006)(Wang,W.,Kong,M.,Du,Z.,Sun,L.,2009),butthe
resultoftheirresearchalsodoesnotgivetherelationship
betweenstairsclimbingabilityandcentroidposition.
In this paper, we take the track mobile robot with front
flipperforexample,justbecausethistypeofrobothasthe
simple mechanical structure and it is well and widely
used in the design and implement of search and rescue
robot. The robot is abstracted as a mass pointplane
model. The process of climbing one stair is analyzed to
build up the mathematic description for the maximal
height the robot could climb over based on the centroid
position of the robot. And we give a precise and
computablesolutionofthisproblem.Itcouldbeabasis
YanGuo,AiguoSong,JiatongBao,HuataoZhangandHongruTang:ResearchonCentroidPositionforStairsClimbingStabilityofSearchandRescueRobot
25
Fig.1.Thepictureofonetypesearchandrescuerobot
Fig.2.Thestructureofrobotsystem
for centroid adjustment to increase the climbing height.
When the robot is running along the nose line of the
stairsafterclimbingthefirststair,thestabilityproblemis
howmuchoftheangletheforwarddirectionoftherobot
deviatingfromthenoselinewillmaketherobotoverturn
back from the side of the robot body. The computable
value of this angle influenced by the centroid position is
discussed. All the works of this paper are to
quantitatively present the centroid position how to
influencethestairsclimbingabilityandstability.
Therestofthispaperisorganizedasfollows:Insection2,
we introduce the search and rescue robot and the robot
model is abstracted. In section 3, we calculate the
relationship between the height that the robot could
overcome and the centroid position. In section 4, the
stability during the stairs climbing is analyzed, based on
the centroid projection method. Section 5 presents our
experimentresults,andsection6concludesthepaperand
suggestsfutureworks.
2.RobotModel
The track mobile robot with front flipper is the widely
usedstructureforthesearchandrescuerobotforitsgood
travelingabilityandreliability.Inthispaper,wetakethis
type robot, shown in Fig.1, for example, which is
designedbyourselves(Guo,Y.,Bao,J.,Song,A.,2009).
This robot has a basic vehicle body and different kind
equipment located on it. The vehicle body and the
equipmenthaveindependencecoordinatedsystemssuch
Fig.3.Themasspointplanemodel
as vehicle body coordinated system OXYZ and
equipment coordinated systems OiXiYiZi which are
showninFig.2.Sothecentroidpositionisactuallyrelated
to individual coordinated system. The vehicle body
coordinated system is selected as the main coordinated
system which all the centroid positions in equipment
coordinated systems should transform to. The centroid
position of the whole robot system should be built up in
thismaincoordinatedsystem.
( )
1
1
n
i
i
n
i
i
M m
M m
=
=
+ +
=
+
,
i i
c c A
C
(1)
In equation (1), M is the mass of the vehicle body,
,
c is
the centroid position of the vehicle. i
m
is the mass of ith
equipment device, i
c
is the centroid position in ith
equipment coordinated system and i
A
is the offset
vector between main and ith coordinated system. The
result C is the centroid position of the whole system in
themaincoordinatedsystem.
The robot system is abstracted as a mass point plane
model with the mass point Centroid
P
that has all the mass
of the system and the vehicle body plane with no mass
but all of the geometric constraints. Shown in Fig.3,
( , , )
Centroid
P d l h = C
and all the arguments will be under
thismodel.
3.ClimbingHeightwithCentroidPosition
In general opinion, the height of the stair that the robot
could overcome is decided by the power of the robot, in
another word, it means that much more power of the
systemcoulddrivetherobottocoverthehigherstair.Itis
right, but not all the time, because the centroid position
also plays much more important role to decide the
maximalheightofcoveringstair.
When the robot is climbing one stair, shown in Fig.4, the
distance between the mass point and the plane is h, and
thedistancebetweenthemasspointandthetailofrobot
is l. Selecting any point S with distance lx, D is the
distancebetweenthetouchpointNandtheprojectionof
pointSonthestairplane,andDisrelatedtoandlx.
InternationalJournalofAdvancedRoboticSystems,Vol.7,No.4(2010)
26
Fig.4.Thepictureofclimbingonestair
( )
cos
sin sin
,
cos
sin sin
x x
x
x x
H H
l l
D l
H H
l l
m
m m
m
m
m m
 
<

\ .
=
 
>

\ .
(2)
During the mass pointplane mode moving to climb the
stair, the touch point N moves to O along the plane, and
theangleincreases.BasedoncurvesshownintheFig.4,
when
sin
x
H
l
m
<
, ( ) ,
x
D l m has the maximum, and when
sin
x
H
l
m
>
, ( ) ,
x
D l m hastheminimum.When
( )
2
,
sin 0
sin
x
x
D l H
l
m
m
m m
c
= =
c
(3)
( ) ,
x
D l m isextremum.So
( )
3
sin , , isextremum
x
x
H
D l
l
m m =
(4)
The critical angle
( )
x
l m'
, which is between the plane of
modeandthegroundwhen ( ) ,
x
D l m isextremum,is
3 arcsin
x
H
l
m
 
' = 

\ .
(5)
The mass point ( ) , ,
Centroid
P d l h has the same kinetic
featurewithitsprojectionP,thecriticalangle c
m'
is
3
arcsin
c
H
l
m
 
' =

\ .
(6)
At this time, the trend of moving to the right is zero for
themasspoint ( ) , ,
Centroid
P d l h .Italsocouldbeunderstood
asthattheprojectionofthemasspoint ( ) , ,
Centroid
P d l h on
thewaterplanewouldnevermovetotherightbutmove
Fig.5.Theclimbingheightbasedoncentroidposition
to the left with the angle
m
going on increasing. At this
time,thegravitylineofthemasspointmusthavemoved
over the touch point N or the gravity line travels across
thetouchpointN(w=0),otherwise,themasspointplane
mode willoverturn back and fail toclimb the stair. So at
thecriticalcondition,
( )
max
,
tan
sin
c
c
H l h
l h m
m
' =
'
(7)
Fromtheequation(6),weknowthat
( )
3
max
2 2
3 3
,
tan
c
H l h
l H
m' =
(8)
Consideringtheequation(7)and(8),
3
max 2
3
max
2 1
3 3
max
( , )
( , ) 0
( , )
h H l h
H l d l l
l H l d
+ =
(9)
8 2 2 2 2
9 3 3 9 9
max max
( , ) ( , ) 0 l H l h h H l h l + = (10)
The equation (10) is the important result of the
relationship between the maximal height the robot could
climb over and the centroid position. If the centroid
position is given before the robot working, we could
calculatethatwhatheightthestairorobstaclewithcould
beclimbedover.Andbeforetherobotstartingtoclimba
stair, the height should be measured and recognized to
contrast with the maximal height. If the maximal height
is larger than that measured, it could safely work,
otherwise,therobotmustterminatetheprocess.Itisvery
meaningful for the safe and successful working of the
search and rescue robot. On the other side, when we
understandthemaximalheighttherobotcoulddealwith,
we must adjust the centroid position based on the result
inequation(10)inadvance.
Thefigure,showninFig.5,givesavisualdescriptionfor
YanGuo,AiguoSong,JiatongBao,HuataoZhangandHongruTang:ResearchonCentroidPositionforStairsClimbingStabilityofSearchandRescueRobot
27
Fig.6.TheprojectionofcentroidonplaneA
Fig.7.TheparametersonplaneA
the equation (10). We could find that the influence of l is
nearly linear with the fixed h, but the tiny increase of h
will lead to the steep decrease of maximal height. So for
theheighttherobotcouldclimb,hismuchmoresensitive
thanl,andthebestwayofenhancingtheperformanceof
climbingheightisdecreasingthehratherthanincreasing
thel.
4.StabilityduringStairClimbing
From the above mentioned discussion, the problem of
turnover back around the button side of the robot
throughfindingthemaximalheightrobotcouldcoverto
avoid climbing the stair higher than that one. When the
robothascaughtthefirststairandgoesonrunningalong
the nose line of stairs, the problem of the robot is
protectingtherobotturnoverbackaroundonesideofits
body because of the flip between each track and the
stairs.
The gravity line projection method is used as the
judgment of the stability and the method is to analyze
whethertheprojectionofthemasspoint Centroid
P
isinthe
projection region of the vehicle body plane on the water
level plane. When the projection of Centroid
P
is in the
projection region of the vehicle body plane, the robot is
withthestability.Thepositionofprojectionof Centroid
P
in
theprojectionregionofthevehiclebodyplaneismoving
with the change of the
m
between nose line plane and
water level plane and the u that between the
Fig.8.TheprojectionfromplaneAtoplaneG
Fig.9.TheparametersonplaneG
directionandthenoseline.Sotheproblemistransferred
tohowmuchoftheangle u theforwarddirectionofthe
robot deviating from the nose line will make the robot
turnover back from the side of the robot body under the
fixedangle
m
.
Shown in Fig.6, G
l
is the projection of the distance from
the mass point to the nose line plane A, and point O is
theprojectionof Centroid
P
onplaneA,and
tan
G
l h m =
(11)
The distance between projection of Centroid
P
and the near
side line of the vehicle body plane on the water level
plane is adopted to describe the stability level of the
system. If the distance equal to zero, that means the
critical stability. Shown in Fig.7, OP is the distance
betweenpointOandnearsideofthevehiclebodyplane,
tan sin
2
D
OP h d m u = + (12)
The right angle between OP and the side could be
divided into two angles 1
u u =
and
2
2
t
u u = . Shown
inFig.8,allthecontentsinplaneAareprojectedtowater
level plane G. So in plane G, shown in Fig.9, angles 1
u'
and 2
u'
are the projections of angel 1
u
and 2
u
.
Accordingtothesolidgeometryangleequation,
InternationalJournalofAdvancedRoboticSystems,Vol.7,No.4(2010)
28
Fig.10.Thecurvesof 2
( , ) f u m
2 2
1
cos( ) 1, 2
2
1 tan ( )cos
2
i
i
i
t
u
t
u m
' = =
+
(13)
1
2 2
tan
arccos( )
2
tan cos
t u
u
u m
' =
+
(14)
2
2 2
1
arccos( )
2
1 tan cos
t
u
u m
' =
+
(15)
And O P ' ' istheprojectionofOPanditequalsto
2 2
1 sin sin O P OP u m ' ' = (16)
Sothedistancebetweentheprojectionof Centroid
P
andthe
nearsidelineofthevehiclebodyplaneonthewaterlevel
planeisM,and
1 2
( , , , ) sin( ) M h d O P u m u u ' ' ' ' = +
(17)
Becauseoftheanglesummationformula,
( )
2 2
2 4 2 2
cos 1 sin sin
( , , , ) tan sin
2
cos sin sin cos
D
M h d h d
m m u
u m m u
m m u u
 
= +

\ . +
(18)
Wedefinedthetwopartsinequation(18)asbellows,
1
( , , , ) tan sin
2
D
f h d h d u m m u = + (19)
( )
( )
2 2
2
2 4 2 2
cos 1 sin sin
,
cos sin sin cos
f
m m u
u m
m m u u
=
+
(20)
The curves of 2
( , ) f u m
are shown in Fig.10, and the
curves demonstrate that 2
( , ) 0 f u m >
except
2
t
m = .
Because the result of function 2
( , ) f u m
is just concern to
thebasicallyzoomofscale,sowecallitthescalefunction
oftheequation(18).
Fig.11.Thefigureof m'
withD=40cm
Observing the function 1
( , , , ) f h d u m
, it domains whether
the value of equation (18) could reach to zero, so it is
calledthebasisfunctionofequation(18).The max
u
which
makes the basis function equal to zero is the critical
constraint and that also means the maximal
stabilitykeepingangle.
Becauseof 1
( , , , ) 0 f h d u m >
,
2
0 , 0 arctan
2 2
2 2
0 , arcsin arctan
2 tan 2 2
D d
h
D d D d
h h
t
m
u
t
m
m
+
s <
+ +
s <
(21)
Considering the equation (21), the angle m'
equal to
2
arctan
2
D d
h
+
isthemaximalanglewithnostabilityloss.
It is shown in Fig.11. When the m m' < , the
max
2
t
u = ,
when m m' > , max
2
arcsin
2 tan
D d
h
u
m
+
=
. So if we know the
centroid position ( ) , ,
Centroid
P d l h , we could calculate the
max
u
andsetthecontrollertomakesure max
u u <
.Andif
we know the performance of the controller of the robot,
weshoulddesignthecentroidpositionbasedonequation
(21)tomatchtheperformanceofthecontroller.
And from the Fig.11, we could know that the tiny
increase of h is much more sensitive on the decrease of
m'
than that of d. So if we want to enhance the stability
during the stairs climbing, the d should be decreased as
muchaspossible.
5.Experiment
We used the search and rescue robot ourselves designed
(Guo, Y., Bao, J., Song, A., 2009) to make a proof of the
method.Thedifferentdevicesareseparatelyloadedon
YanGuo,AiguoSong,JiatongBao,HuataoZhangandHongruTang:ResearchonCentroidPositionforStairsClimbingStabilityofSearchandRescueRobot
29
Fig.12.Theexperimentofmaximalheightcouldclimb
Fig.13.Theexperimentof max
u
NO.
l
cm
h
cm
Experiment
Hcm
Computation
Hcm
1 30 20 8.3 8.34
2 30 30 5.3 5.37
3 35 35 6.2 6.27
4 35 40 5.2 5.26
5 40 45 6.1 6.14
6 40 20 14.1 14.14
7 45 30 12.5 12.51
8 45 35 10.7 10.74
9 50 40 11.6 11.59
10 50 20 20.6 20.62
Table1.Recordofmaximalheight
the robot to receive the different centroid positions. And
we adjusted the height of stair and drove the robot
climbing it, shown in Fig. 12. The maximal height is
recordedinTable1.
With the different centroid position, the robot is
requested running on the stairs shown in Fig. 13
with 30 , 45 ,60 m = . And the results of max
u
are
recordedinTable2Table3andTable4.
Fromallthetables,thecontrastsbetweentheexperiment
results and the computed results are almost equal. And
we have enough reasons to believe that the analysis and
computemethodofthemaximalheighttherobotcould
NO.
d
cm
h
cm
Experiment
max
u
Computation
max
u
1 0 30 90 90
2 0 40 41 40.89
3 4 35 38 38.37
4 4 45 31 31.63
5 8 30 35 34.72
6 8 40 27 27.46
7 12 35 21 21.60
8 12 45 17 17.11
9 16 35 11 11.20
10 16 45 9 8.75
Table2.Recordof max
u
, 30 m =
NO.
d
cm
h
cm
Experiment
max
u
Computation
max
u
1 0 30 34 33.69
2 0 40 26 26.57
3 4 35 24 24.57
4 4 45 20 19.57
5 8 30 22 21.80
6 8 40 16 16.70
7 12 35 12 12.88
8 12 45 10 10.08
9 16 35 6 6.52
10 16 45 5 5.08
Table3.Recordof max
u
, 45 m =
NO.
d
cm
h
cm
Experiment
max
u
Computation
max
u
1 0 30 21 21.05
2 0 40 16 16.10
3 4 35 15 14.79
4 4 45 11 11.60
5 8 30 13 13.00
6 8 40 10 9.83
7 12 35 7 7.52
8 12 45 5 5.86
9 16 35 3 3.78
10 16 45 2 2.94
Table4.Recordof max
u
, 60 m =
climboverandthemaximalstabilitykeepingangle max
u
arejustifiableandcorrect.
6.Conclusion
In this paper, we are trying to find out the relationship
between the stability of stairs climbing and the centroid
position of the search and rescue. Considering the work
situation of the search and rescue robot, the system is
abstractedasamasspointplanemodel.Weanalyzeand
calculate the process of climbing one stair, and the
relationship between the maximal height the robot could
climb over and the centroid position is developed
InternationalJournalofAdvancedRoboticSystems,Vol.7,No.4(2010)
30
through making the robot never turnover back around
the bottom side. The mathematic description is given in
this paper. And we also give an analysis of the stability
during the stairs climbing. In order to protect the robot
not to turnover back around one side of its body, we are
tryingtofindandcalculatethemaximalstabilitykeeping
angle deviating from the nose line when the robot could
keep its stability. And a precise and computable solution
of this problem is given in the paper. The experiment
using the real search and rescue robot demonstrates its
justificationandcorrection.
Alltheresultofresearchisbasedonthephysicalfeatures
of the search and rescue robot, and it hopes to find
another way instead of adding more electric devices to
enhance the performance of the robot. Our purpose is
exploring the potential ability of the robot mechanical
structure and keeps the control and electric systems
concise and trustworthy. In future, we should do the job
of decreasing the weight of the control and electric
systems based on physical optimization such as centroid
positionandotherslikethat.
7.Acknowledge
This research is made possible with support from the
Project under Science Innovation Program of Chinese
EducationMinistry(No.708045).
8.References
Ko, A., Lau, H. (2009). Intelligent Robotassisted
Humanitarian Search and Rescue System.
International Journal of Advanced Robotic Systems,
Vol.6,No.2,pp.121128,2009
Guo, Y., Song, A., Bao, J., Tang H., Cui, J. (2009). A
CombinationofTerrainPredictionandCorrectionfor
Search and Rescue Robot Autonomous Navigation.
International Journal of Advanced Robotic and
Systems,Vol.6,No.3,pp.207214,2009
Kim, B. H. (2009). Centroidbased analysis of
Quadrupedrobot Walking Balance. International
ConferenceonAdvancedRobotics.Munich,Germany,
2009
Zhu, X. L., Wang, H. G., Fang, L. J., Zhao, M. Y., Wang,
L.D. (2006). Centroid Adjustment Control on
Autonomous Obstacle Negotiating Inspection Robot.
Robot,Vol.28,No.4,pp.385388,2006
Kim,B.S.,Vu,Q.S.,Song,J.B.,Yim,C.H.(2010).Novel
Design of a Small Field Robot with Multiactive
Crawlers Capable of Autonomous Stair Climbing.
JournalofMechanicalScienceandTechnology,Vol.24,
No.1,pp.343350,2010
Wang, W., Kong, M., Du, Z., Sun, L. (2009). Effect and
Motion Planning of Manipulator for Obstacle
Negotiation of Track Robot. Journal of Mechanical
Engineering,Vol.45,No.8,pp.617,2009
Liu, J., Wang, Y., Ma, S., Li, B. (2005). Analysis of Stairs
Climbing Ability for a Tracked Reconfigurable
Modular Robot. IEEE International Workshop on
Safety, Security and Rescue Robotics. Kobe, Japan,
2005
Pinhas, B. T., Ito, S., Goldenberg, A. (2009). A Mobile
RobotwithAutonomousClimbingandDescendingof
Stairs.Robotica,Vol.27,No.2,pp.171188,2009
Liu,Y.,Liu,G.(2009).TrackstairInteractionAnalysisand
Online Tipover Prediction for a Selfreconfigurable
Tracked Mobile Robot Climbing Stairs. IEEE/ASME
Transactions on Mechatronics, Vol. 14, No. 5, pp.
528538,2009
Woo,C.K.,Choi,H.D.,Yoon,S.,Kim,S.H.,Yoon,K.K.
(2007). Optimal Design of a New Wheeled Mobile
Robot Based on a Kinetic Analysis of the Stair
Climbing States. Journal of Intelligent and Robotic
Systems,Vol.49,No.4,pp.325354,2007
Guo, Y., Bao, J., Song, A. (2009). Designed and
Implementation of Semiautonomous Search and
Rescue Robot. IEEE Conference on Mechatrinics and
Automation.Changchun,China,2009