Industrial Robot: An International Journal

Design and manufacturing of a tele-operative rescue robot with a novel track arrangement
Nima Enayati Farid Najafi
Article information:
To cite this document:
Nima Enayati Farid Najafi, (2011),"Design and manufacturing of a tele-operative rescue robot with a novel track arrangement",
Industrial Robot: An International Journal, Vol. 38 Iss 5 pp. 476 - 485
Permanent link to this document:
[Link]
Downloaded on: 01 May 2016, At: 07:06 (PT)
References: this document contains references to 14 other documents.
To copy this document: permissions@[Link]
The fulltext of this document has been downloaded 344 times since 2011*
Users who downloaded this article also downloaded:
(2012),"The effect of an audit firm's brand on security pricing", International Journal of Emerging Markets, Vol. 7 Iss 4 pp.
430-442 [Link]
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

(2013),"Turning crisis into opportunity: The key role of leadership", Strategic Direction, Vol. 29 Iss 4 pp. 29-32 http://
[Link]/10.1108/02580541311311302
(2014),"The relationship between enduring leadership and organizational performance", Leadership & Organization
Development Journal, Vol. 35 Iss 7 pp. 590-604 [Link]

Access to this document was granted through an Emerald subscription provided by emerald-srm:331053 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please visit
[Link]/authors for more information.
About Emerald [Link]
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication Ethics
(COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

 Research article

Design and manufacturing of a tele-operative

rescue robot with a novel track arrangement
Nima Enayati and Farid Najafi
Department of Mechanical Engineering, K.N. Toosi University of Technology, Tehran, Iran

Abstract
Purpose – The paper aims to describe the design and manufacturing process of a tele-operative urban search and rescue tracked robot and discuss the
advantages of a proposed novel track arrangement and other additional mechanisms, which help the robot to gain high manoeuverability on rough
terrains.
Design/methodology/approach – Using a simplified static model, required torques are calculated and appropriate mechanisms and geometric
dimensions are chosen. Next, stress distribution is analyzed in the parts, deploying both classic and numerical methods and to complete the procedure
parts are fabricated and assembled together. The architecture of control system and the user interface is introduced. Finally, the robot is tested on a
standard test arena and the results are compared with another similar search robot.
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

Findings – A tele-operated rescue robot with considerable capabilities is designed and manufactured. The novel track arrangement and new rear arm’s
mechanism are tested and compared with a robot in the same class and higher performance is achieved on the evaluation.
Originality/value – Although the implemented locomotion mechanism is the common tracked type, adding the center tracks and arrangement of arms
are original ideas which help the robot to gain high manoeuverability. The proposed rear arm’s linkage mechanism generates a limited rotational path
and has an acceptable strength for a robot working in rescue missions.

Keywords Search and rescue, Industrial robotics, Robots, Mechanisms, Teleoperation

Paper type Research paper

Introduction is present, some form of direct operating interface will be

required. Even when full autonomy becomes possible,
Mobile robot technology has reached a level of maturity at alternative forms of tele-operation will be needed in emergency
which it should soon be possible to use rugged, portable situations. Since the rescue sites and rescue activities are complex
mobile robots for urban reconnaissance missions that include and rescue robots need to cross over different obstacles to reach
indoor and outdoor traverses and onboard mapping. Mobile destination, they should be rugged, and have all-weather
robots can be deployed in environments where it may not be platforms capable of maneuvering in desert or beach sand,
safe or affordable to send humans. snow, grassy or wooded terrain, and inside buildings. They
Research on the mobile robot for medical service, urban search should also be small, light, portable and capable of climbing
and rescue (USAR), anti-terrorist and street battle, has attached stairs and sustaining falls. Because of advantages such as
a lot attention. One particularly, challenging field for tele- excellent stability and low terrain pressure, more and more
operated mobile robots is USAR. In the future, the first researchers take interest in tracked mobile robot which can move
responders on the scene of an earthquake or other disasters will in unstructured environment quickly and the development of
deploy a group of autonomous robots. These robots co-ordinate tracked robots with different functions has been carried out
amongst themselves, search the area, detect and identify victims, extensively. The “TALON” robots, built by US Company
report their conditions, deliver first aid and assist human rescuers Foster-Miller, have already performed over 10,000 explosive
when they arrive. While such dreams are a long-term approach, ordnance disposal missions. “SWARDS” is another robot used
technologies are already being developed that work towards this in the field. The iRobot “Warrior 700” can carries payloads
goal. In the meantime, USAR robots must have at least partial exceeding 68 kg, travels over rough terrain and climbs stairs while
tele-operation. Different degrees of autonomy are starting to performing a variety of critical missions ([Link]/sp.
become possible for mobile robots, however, until full autonomy cfm?pageid¼172 (accessed 18 November 2010)).
In this paper, an illustrative description of the Adora project
The current issue and full text archive of this journal is available at at K.N. Toosi University of Technology is presented (Figure 1).
[Link]/[Link] In the first section, innovative mechanisms and mechanical
design procedure are highlighted. Diverse locomotion systems

Industrial Robot: An International Journal

38/5 (2011) 476– 485 This work has involved the efforts of other members of Pars robotic group
q Emerald Group Publishing Limited [ISSN 0143-991X] whom the authors would like to thank: Gholamhosein Mahaseni,
[DOI 10.1108/01439911111154045] Mahdi Falahinejad and Siavash Malektaji.

476
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 1 Adora in different situations

(a) (b)
Notes: (a) Maneuvering on rough terrain; (b) climbing stairs

used in mobile robots and their capabilities are studied and a special joint mechanism (Masayuki et al., 2004) and KOHGA
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

new mechanism is developed to enhance the manoeuverability consists of eight units and each unit is connected either by
of a mobile robot on uneven terrains. Second section presents 2-DOF (degrees of freedom) active joints or 3-DOF passive
the stability and mobility analyses from statics and dynamics, joints, so that the robot can have both the capability to climb
focusing on worst situations in motor selection. In the third over obstacles as well as adaptation to irregular surfaces
section, the procedures of stress analysis for the mechanical (Kamegawa et al., 2005).
parts is discussed. Control system of the robot and electronic Because of advantages such as excellent stability, simple
hardware are described in the last section. control system, good manoeuverability, rapid zero radius
turning and low terrain pressure, the tracked type has been
Mechanical design widely applied in uneven or unstructured environment
(Bekker, 1969).
Locomotion mechanism Toin Pelican series of robots (Koyanagi et al., 2004) built by
A wide variety of different designs have been used in USAR Toin University of Yokohama for urban search are one of the
robots, and each has both pros and cons. Among these well-known robots in Robocup competitions. These robots
models, three types are more tried: leg type, wheel type and have a neat design and they employ two pairs of small tracked
tracked type.
arms to climb obstacles. These arms efficiently help the
Wheels are cheaper and have good performance on flat
robots on uneven terrains; however, their not-very-solid
areas. For example, robot FUMA travels around with four
design can be damaged on rough arenas. They also mounted a
wheels and uses a specially made camera arm to lift up and
precise 5-DOF manipulator on one of their robots and
withstand the weight of the robot (Shiroma et al., 2004). But
successfully manipulated some objects in RoboCup 2010 in
when it comes to the mobility on rough terrains, wheels are
Singapore.
not useful anymore. As a matter of fact, wheels usually can
Redback and CASTER robots (Sheh, 2006; Kadous et al.,
climb up only obstacles which are lower than the center of the
2005) designed by the ARC Centre of Excellence for
wheel. Thus, their capability is proportional to their wheel’s
diameter, while a tracked or snake-like robot can achieve the Autonomous Systems of the University of New South
same ability in a considerable smaller size. Legged robots have Wales, Sydney are low-cost robots for education and
a flexible platform and can travel across uneven terrain quite research. Both of them use tracks for locomotion. The
well (Shiroma et al., 2004); however, they need relatively high Redback is a platform based on a toy vehicle which can climb
numbers of actuators and sensors which results in stairs and overcome many obstacles using its two pairs of
complicated dynamic modeling and expensive control expandable tracks. CASTER has two pairs of rubber tracks
system. A combination of wheeled locomotion and passive for locomotion. The front pair follows a triangular path,
adaptation have been used in SOLERO (Michaud et al., allowing the robot to climb over obstacles and stairs whilst a
2002) which shows good climbing abilities without any conventional flat rear pair provides additional traction.
specific active control. This mechanism has better efficiency Mechatronics Research Laboratory rescue robot group from
compared to active designs such as legged rovers, but it is Azad University of Qazvin used the same triangular tracked
more suitable for planetary exploration rather than urban arm in their NAJI-V ([Link]/2006/TDP/%
search, duo to its poor performance in stairs climbing. 28IRAN%29_MRL_TDP.pdf (accessed 15 November 2010))
Some other different designs have been proposed for USAR robot and added the ability to rotate the arm, which
robots like hopping robots, snake-like robots and UAVs. substantially enhanced the manoeuverability of the design.
Many variations of snake-like robots have been developed Stairs climbing is a crucial performance for urban operation
during recent years. For example, the ACM is an intelligent, robot. So, the design should satisfy requirements to climb all,
biologically inspired mobile robot (Hirose, 1993), Souryu I-V kind of stairways. One of the design parameters that should be
robots are constructed by connecting three crawler units using considered is the total spanning length, the track length plus

477
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

the expandable tracks length, which must be larger than three Figure 3(c). Placing both front and rear arms in a plan with
stair steps. This allows the robot to travel on a stair smoothly. main tracks would solve the gaps problem. This would
The stairs dimension includes the riser height and the tread decrease the width of the mechanism which seems an
width. The riser height h, which relates to human step uplift, advantage, but it also causes instability and increases the flip-
changes from 120 to 180 mm. The tread width b, which over risk. It should be mentioned that although smaller robots
relates to human foot length, varies from 200 to 350 mm (Liu can explore areas like holes or narrow passages easier, stability
et al., 2005). It is important to design the robot’s dimension is a more important characteristic and should not be
reasonably as the stairs dimension is fixed in a range. So, the diminished by the size of the robot. To improve stability,
total spanning length should be larger than 790 mm. In the width should be increased, but this will also extend the
addition, the front arms should be longer than 180 mm to untracked space beneath the robot, which lessens robot
make a reliable contact point with the first stair step. manoeuverability on rough terrains.
Based on preliminary models, an innovative design is
Arms proposed to satisfy all design requirements. The conventional
In order to enhance maneuvering ability, a pair of front
front pair of extended tracks can rotate 3608, allowing the
flippers has been widely used in explorer robots (Matthies
robot to climb over obstacles and stairs while a rear pair
et al., 2000; Moosavian et al., 2009). Front flippers, help in
parallel to the main tracks provides additional traction and
climbing up stairs and extend the robot length for stepping
weight transfer Figure 4. Thus, while the rear arms will not
over ditches. In addition, the expandable tracks can be folded
hinder movement, the stability remains high.
to reduce the size of the robot and achieve smaller turning
The robot also can climb bigger obstacles by using front
radius in narrow passages. Adding another pair of arms at the
rear of the robot allows it to transfer weight to the front, which flippers in reverse direction. Figure 5 shows the procedure of
will increase the traction force on the contact point with the climbing up obstacles from the robot rear. First, front flippers
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

obstacle (Figure 2). rotate to the rear to support the body, and then the robot moves
The arrangement of the robot’s arms affects its size, stability toward the obstacle. When the body touches the obstacle, the
and traverse capabilities, which are cardinal factors in front flippers rotate clockwise to raise the body and move
mobility of a USAR robot. Figure 3 shows three different the center of gravity (CG) over the obstacle. This ability lets the
designs introduced in literature for tracked robots with four robot to climb obstacles up to a height of 30 cm.
expandable tracks. Rear arms are placed parallel to the side track, thus they
As shown in Figure 3(a) and (b) in robots with four flippers, have to rotate around the rear wheels of the side tracks. To
there is a gap between the front and rear flippers, which acquire the desired motion path, a four-link slider mechanism
cause their main disadvantage. External objects can readily is designed (Figure 6). The torque T transmitted from the
stick in these gaps and hinder movement of the robot. motor rotates the link one of the mechanism. One end of the
Especially, when the gaps are out of the vision of camera, the second link is mounted on a slider bearing and the other end
operator will not see the obstacles and he/she has to rotate carries the sprockets and tracks.
camera to be able to release the robot. Hence, it takes much Power is transmitted to the arms by means of a timing
more time to traverse uneven terrains. In order to alleviate chain. However, because of oval rotation, the distance
this problem and reducing width of the robot Resquake between the rear wheel of the main track and that of rear
(Moosavian et al., 2009) used an improved design, shown in arm will not remain constant. But the length of the chain is

Figure 2 Rear arms help the robot to climb obstacle Figure 4 Proposed tracks arrangement

Figure 3 Preliminary arrangements of arms used for USAR robots

(a) (b) (c)

478
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 5 Climbing up obstacle from rear

(a) (b) (c)

Figure 6 Rear arms motion path make the robot less stable. On the other hand, nearing the two
side tracks will reduce the needed couple in zero radius
T Slider turning. In order to compromise between the two
contradictory aspects of providing a small robot with high
stability, a pair of central tracks has been developed for Adora.
L2
These extra tracks ensure that obstacles under the robot
L1
would not hinder its mobility (Figure 4).
Both body and side tracks in each side of the robot are
powered by a powerful DC motor. However, sprockets used
for body tracks are half the size of that of side ones. Hence,
body tracks exert two times more traction force and they have
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

half the velocity of side tracks. Higher traction force helps the
robot to traverse rough rubbles easier and guarantees
continuous movement on uneven terrains. To prevent
excessive friction caused by lower speed of central tracks,
they are designed to not to have contact with the surface on
flat areas. Therefore, the robot can achieve its maximum
speed on a flat terrain.
Designed tracks are assembled from timing chain and latex
pieces. Using metal chain provides sufficient tensile strength
and allows flexibility to change the track length. In addition to
constant. To compensate the difference of the distance, a high friction of the latex pieces, its compliance structure gives
spring mechanism is placed between the arm and the side large area of ground contact, which results in even higher
track, as it is shown in Figure 7. friction force. Furthermore, the latex pieces can damp
A 1-DOF arm with a pan-tilt unit is also designed to carry the moderate shocks to the system caused by collisions.
sensor head and place it as close as possible to the target. Using
two DC servo motors, the pan-tilt mechanism enables the
operator to rotate the sensor head and to look around. The arm Static analysis
is 440 mm long and has a span of 160 degrees of rotation. CG of the system
In static obstacle performance analysis, the CG of the robot
Tracks plays a very important role. The system’s CG is changed with
As mentioned in previous part decreasing the width of the the motion of expandable tracks and the sensor head arm, and
robot is desirable in exploring narrow passages, but it will also the stability and obstacle performance are changed accordingly.
The CG position is also needed in different conditions in order
Figure 7 Rear arm and its rotational spring-chain mechanism
to calculate external forces exerted to the robot.
The co-ordinate frame is fixed to the center of the rear
sprocket of the main track (Figure 8). The system’s mass
consists of four parts, robot body, front arms, rear arm and
the sensor head arm, which are represented by m0, m1, m2
Chain-tightener and m4, respectively. According to the CG equation, the
mechanism

Figure 8 CG position of the robot’s mass system

m3g
q3
m2g
Y L3 m 1g
Link 2 q2 C.G d
X q1
L2 m 0g
L1
Link 1
L0

479
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

system’s CG position (Lx, Ly) in the robot co-ordinate system acceleration of the system and rotational acceleration of the
XOY can be obtained as: wheels could be assumed to be zero. Substituting ax ¼ 0 in
8 the Newton’s second rule and assuming the weight of the
>
> LX ¼ m0 ðL0 =2Þ þ m1 ðL1 cos½u1 þ L0 Þ þ m2 L2 cos u2 robot to be 45 kg, the minimum traction force is calculated as:
>
> .
<
þ m3 ðL3 cos½u3 þ L0 Þ

m0 þ m1 þ m2 þ m3 F t ¼ 312:2 N
ð1Þ
>
>
>
> m1 L1 sin u1 þ m2 L2 sin u2 þ m3 ðL3 sin½u3 þ dÞ

: LY ¼ Substituting the wheel radius in the moment equation with
m0 þ m1 þ m2 þ m3 respect to the center of the wheel and assuming zero
In this system, m0 ¼ 29.3 kg, m1 ¼ 7.4 kg, m2 ¼ 6.1 kg, acceleration, the total required torque for traction is obtained.
m3 ¼ 1.2 kg, L0 ¼ 0.424 m, L1 ¼ 0.053 m, L2 ¼ 0.210 m, Dividing the calculated Tt by 2 gives the desired torque of
L3 ¼ 0.439 m and d ¼ 0.191 m. It should be mentioned that each locomotion motor, Tm:
all masses are calculated using SolidWorks based on used T m ¼ 10:9 N:m
material’s density.
For convenience, it is assumed that rear arm is a one joint In order to calculate the desired torque for arms, the worst
arm. Also, the CG of the body assumed to be at the geometric situation which is shown in Figure 10 is considered. In this
center O0. case, the whole weight of the robot is supported by the arms.
Letting u1, u1 ¼ 0 and u3 ¼ 45 degrees in equation (1), the
Rear arm’s mechanism horizontal position of the robot’s CG is computed as:
In order to calculate forces applying to the rear arm’s
mechanism, the free body diagram of the linkage is shown in l X ¼ 0:262 m
Figure 9. Although there exists a friction force F2x, it is negligible
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

compared to the vertical force applied to the arm. Therefore, the Solving the moment equations, the force exerted on each arm
only horizontal reaction in the L2 link F1x is assumed to be zero. is obtained as:
Writing the equation of moment about point O1: F f ¼ 286:7 N F r ¼ 184:8 N
X T
M O ¼ 0 ! F 1y ¼ ð2Þ The torque of the front arm’s motor is calculated easily as:
d 1 cos a
T f ¼ 71:7 N:m
Writing the equation of moment about point O2:
F 1d2 For the rear arms, equation (2) is used to calculate the torque,
F2 ¼ ð3Þ by letting a ¼ 0 and F2 ¼ 184.8:
l
Substituting equation (2) in equation (3), we get: T r ¼ 80:0 N:m

Td 2 The locomotion system consists of two separate sides

F2 ¼ ð4Þ connected to the main body and the central tracks, while a
ld 1 cos a
motor independently actuates the track at each side. Flippers
where F2 is the normal force exerted to the arm by the surface are coupled together and powered by a motor, so are the rear
and d1 ¼ 0.16 m, d2 ¼ 0.08 m and l ¼ 0.23 m. arms. Assuming the power losses due to friction in power
The dynamic model of the system can be obtained by using transmission components to be less than 30 percent of the
kinematics analysis and applying Lagrange’s equations for a maximum power, 14 N m output torque would be sufficient
mobile robot. Here, the torque needed for propelling the for each side. The choice ends up with 24 V DC brushless
robot on ramps and stairs has to be calculated. Therefore, to 110 W motors for both locomotion system and arms. The
avoid cumbrous analysis, it is practical to obtain the motors output velocity is 3,600 rpm and it can produce a
maximum torques assumed to be applicable to the system. maximum torque of 0.37 N m. A gear box with a ratio of 40:1
This occurs when the robot is climbing up a ramp. Therefore, is deployed to increase the motor torque to 14.8 N m which
considering a simplified model, the minimum torque needed satisfies the requirement. Choosing the wheel diameter of
for climbing a 45 degree ramp is calculated. The velocity of 130 mm, the maximum speed of the robot is calculated to be
the robot is assumed to be constant. Thus, the linear 0.61 m/s.
Figure 9 Free body diagram of the rear arm’s linkage mechanism

o2
α q d2 F3
F1y
d1 1 F1y F1x
o1 L1 L2
F1x
T F1x
F1y

F2

480
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 10 Weight of the robot supported by arms over a ditch Following the stress analysis of all mechanical parts and
redesigning of those required, parts are fabricated using
l2 l1 l3 common methods like water jet cut for flat parts, lathe for
rotatory parts and CNC machines for high precision parts.
The last step is to assemble controllers and other electronic
parts, which is discussed in the next section.

Electrical design and control system

As it is shown in Figure 13, the control system of the robot
consists of various parts:
.
An operator station. A laptop (VAIO VGN-FE690P with
Wireless LAN 802.11 a/b/g, 2 GB RAM, CPU a.86
The speed of the arms should be less than 15 rpm and they Centrino Duo).
need higher torque than locomotion system. Therefore, a .
Controller circuits. A main board, sub-controllers for
combination of a gearbox with a ratio of 100:1, and a helical sensors and motors.
gear set with a ratio of 2.56 is deployed. Helical gears have .
Sensors and actuators. Pyrometer, CO2 sensor,
higher strength comparing to spur gears and also produce less potentiometer, encoder, tilt sensor, microphone and
noise (Shigley and Mischke, 2003). This gear sets results in a camera.
torque of 94.7 N m and a rotational velocity of 14 rpm for
arms’ shafts. The main board on the robot deploys two 8-bit ATMega 128
microprocessor and a highly integrated Ethernet controller
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

(Realtek RTL8019AS) as a module to convert usual types of

Stress analysis data (in serial or parallel) to TCP/IP standard data. The
The robot should be resistant to hardware failure. If a robot information exchange between the robot and operator station
breaks down in any way during a stressful USAR mission, it is carried out through wireless LAN by means of a wireless
may be impossible to repair it in a realistic time frame, and router, mounted on the robot. Two sub-controllers and four
lives may be lost because of it (Cavallin and Svensson, 2009). BLDC motor drives are connected to the main board and
The robot must be very reliable and robust; therefore all parts exchange the speed data of BLDC motors, the position data
should be analyzed to have enough strength before of the arms and various sensors information. This control
manufacturing. On the other hand, analyzing stress structure enables a highly responsive control by distributing
distribution in parts helps to reduce the total weight of the the computation load of the total system among sub-
robot by preventing overdesign. controllers.
Almost all parts of the robot are designed and analyzed in Sensors used in the control system are categorized into two
SolidWorks software (Figure 11). The stress in all shafts and groups, victim-detection sensors and position sensors. Two
simple parts is calculated and by using classic methods of encoders mounted on drive wheels of the tracks, and three
failure like distortion energy (Shigley and Mischke, 2003), the potentiometers coupled to front, rear and sensor head arms,
diameters of shafts are calculated. After designing other parts provide feedback data to form a closed-loop position control
and checking available bearings in market, dimensions are system. A CO2 sensor and a non-contact temperature sensor
revised and the final diameters are chosen. Since classic are mounted to measure victims’ vital conditions. Data
methods are not applicable for some parts due to complexity provided by these sensors are translated through an interface
of the geometry. FEM method is deployed to analyze their circuit and sent to operator station.
strength, using COSMOS Works software. The critical parts High-power lithium-ion polymer rechargeable batteries
are redesigned to achieve a factor of safety of 2-3 (Figure 12). (24 V, 10 A H) are deployed as the onboard energy sources
to extend the energy availability and increase run time.
Figure 11 The complete model in SolidWorks with 731 parts
User interface
Human-robot interaction is a research field concerned with
the interactions between a human user and a robot. This
subject is an important part of USAR robotics because of the
need of the operator to get a good overview of the situation,
and the need to effectively operate the robot during stressful
conditions (Cavallin and Svensson, 2009). Usually, the
human (or “master”) directs the robot (or “slave”) via some
sort of workstation interface located out of viewing distance
from the robot. The human operator is required to have a user
interface (UI) consisting of a display and a control
mechanism, and the robot is required to have power,
effectors and sensors.
As it is shown in Figure 14, the most prominent feature
of UI for the Adora robot is that most of the space is taken up
by the video feed from its main camera. This camera is
installed on a spring attached rod at rear of the robot facing

481
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 12 Flipper’s connection link redesigned to gain a F.S. of 2.3

Model name: Study__teleoperative v4_arms_front_flat bar
Study name: Study 1
Plot type: Static nodal stress stress1
Deformation scale: 5.63895

Von mises (N/m^2)

3.271e+008
2.999e+008
2.726e+008
2.454e+008
2.182e+008
1.910e+008
1.638e+008
1.365e+008
1.093e+008
8.209e+007
5.486e+007
2.764e+007
4.197e+005
Yield strength: 7.500e+008
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

Figure 13 Block diagram of the control system level and the view of the main camera is not informative. In
this case, the user can rotate the arm to elevate the pan-tilt
Operation station unit and place it as near as possible to the victim. An auxiliary
rear view camera is also placed on the robot for backward
Interface
motion. The operator can change the main camera view to the
Joystick Wireless
TCP/IP rear one by pressing a key on the joystick. Also displayed
Keyboard under the main video feed are the various sensor gauges. The
Remote computer
most distinct of these is the virtual 3D model in the lower left
WLAN corner of the screen. The virtual 3D model of the robot
provides the user with information regarding the robot’s
BLDC
controller current tilt and arms’ position. This is useful for complex
Sensor’s arm motor situations where it is hard for the user to orient themselves by
Wireless Camera 1
BLDC lan just looking at the video feed. In the lower right part data
controller about the robot’s internal status are displayed. This data
Video
Front arms motor server include things such as speed, remaining battery level and CO2
Camera 2
BLDC and heat sensors threshold.
controller
The robot’s locomotion, arm’s rotation and the pan-tilt
Rear arms motor
camera are all controlled by a joystick and the key setting is
BLDC
controller designed to be simple, relevant and familiar to create a fast
Right motor Main board
learning UI.
BLDC Micro Other specifications of the robot are listed in Table I.
controller phone

Left motor
Conclusion and results
Interface
circuit Owing to complex environment and involving many factors,
evaluating the performance of a search robot is not an easy
Encoders
Sub controller task. One of the most practical test arenas is a performance
for sensors metrics published by National Institute of Standards and
Potentiometers
Technology (NIST) of the USA, which is also used in
Temp sensor Co2 sensor Tilt sensor
Robocup Rescue Robot League the competitions. The goal of
Adora this competition is to increase awareness of the challenges
involved in USAR applications, provide objective
performance evaluations of mobile robot implementations in
the front-downward view to generate a semi-third-person view complex yet repeatable environments, and promote
of the robot in environment. In addition to the main camera, collaboration between researchers. Winning teams must
two other cameras are also used. At the upper left corner of reliably perform seven to ten missions of 15-30 min from
the interface, a pan-tilt camera view is placed. This pan-tilt various start points to find the most victims.
unit, which includes the camera and victim-detection sensors, Implementing the NIST’s performance metrics and test
is mounted on a 40 cm arm and can be operated arena standards, a test arena was built at K.N. Toosi University
independently of the robot itself. This is useful for of Technology including step field pallets, 45 degrees ramp,
situations where the victim is found to be on an elevated stairs and elevated floor with rollers (Figures 14 and 15).

482
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 14 The UI of the robot

Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

Table I Specification of Adora rescue robot Description of these components can be found at web site: www.
[Link]/projects/USAR/ (accessed 18 November 2010).
Weight (battery included) 42 kg The difficulty levels of the step fields increases form the first one
Width (flippers included) 520 mm to the third one. Since no documentation about the performance
Minimum height 358 mm of other rescue robots could be found in the literature, the
Length (expanded arms) 1,110 mm performance of the robot is compared with another rescue robot
Length (folded arms) 776 mm designed by Pars robotic group in 2006, named PARS II
Maximum velocity 0.61 m/s (Figure 16). This robot has an acceptable manoeuverability and
Maximum angle of ramp climbing Up to 458 had achieved some awards at national Robocup competitions
Number of unique parts 124 ([Link]/projects/ (accessed 10 November 2010)).
Total number of components 731 Lack of rear arms and center tracks makes the robot PARS II a
Time of operation 2h suitable choice for evaluating the improvement gained by adding
the new mechanisms to robot Adora.

Figure 15 The designed test arena based on NIST’s performance metrics and test arena standards

483
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

Figure 16 Robot PARS II were conducted. The average T and the average of the
absolute deviations D of measured times for each mission is
calculated and the results are shown in Table II. As it can be
derived from the results, in most cases Adora has
accomplished the mission in a lower mean time, which
represents the higher manoeuverability of this robot. In
addition while the robot PARS II failed to pass the third level
step field and three-roller elevated floor, Adora has
successfully passed over all standard obstacles. Comparing
the total average time for the trained and beginner driver may
lead to another interesting result. As it was mentioned before,
the UI should be designed in a way that USAR workers be
able to learn how to control the robot with the least possible
training period. Obviously, a beginner operator spends more
time to finish a mission, since he or she might have difficulties
in interpreting the sensory data to conceive the robot’s exact
status and decide on next commands. Therefore, as much as
the mission time for a beginner operator gets closer to that of
a trained one, effectiveness of the UI and simplicity of the
operating method are improved more. The ratio of the total
mission time for the beginner driver to that of the trained one
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

The goal of the test mission is to evaluate the mobility and is calculated as 0.89 for Adora and 0.81 for PARS II, which
manoeuverability of the robot, hence climbing and passing shows that the beginner driver has learned to work with adora
obstacles in different approximations of “real world” disaster faster than PARS II.
scenarios have to be considered. On the other hand, time is a In order to provide direct comparison of robotic approaches
highly valuable commodity in rescue scenarios. A fast search in a public proving ground, Adora was participated in regional
is vital, because of the fact that the lack of food, water and and international Robocup competitions as one of the Pars
medical treatment causes the likelihood of finding a person group’s robots and gained considerable success. The robot
alive to be greatly diminished after 48 h has passed since the successfully passed all the obstacles in competitions’ arenas.
incident. Therefore, how fast the robot can pass the obstacle In addition, during all test missions and competitions robot
is another important factor which has to be taken into Adora flipped over only once, which was less than most of the
account. tracked rescue robots. This shows the new arrangement of
The experience of driver directly influences the test tracks which is proposed here and robot’s width to length ratio
performance. Since the operator generally cannot see the to be effective on gaining high stability on uneven surfaces.
robot directly, he or she has to rely only on the collected data Rear arms were proved to be useful, especially in stairs
by the robot about its nearby environment and decide on the climbing and climbing obstacles.
commands which have to be sent to the robot. This The results in Robocup are based on the number of victims
interaction usually needs practice. Experienced operators are found by both a tele-operated and an autonomous robot and
able to use the robots more efficiently. This allows for a faster the quality of a generated 2D map of the arena which contains
search of an area and as a result more victims can be found. the place of the founded victims. Despite the fact that the
To consider the effect of the experienced driver, each test was short camera arm of Adora prevents it from scoring victims
performed by two people, an experienced driver and an placed over a height of 1.1 m and without achieving the map
amateur driver who had driven the robot for less than 2 h score (the team had only one laser range finder which was
before the tests. mounted on the autonomous robot), the challenging
The test method comprises six missions and the time manoeuverability of the robot helped Pars robotic group to
duration to complete the mission is measured. Missions are gain the third best team award in rescue robot league at the
simply defined by climbing and passing obstacles. All missions Robocup Iran-Open 2010 competitions. The team also was
were performed for five times and by two different operators selected among the ten pre-final teams at the World Robocup
via remote tele-operations. Therefore, for each robot 60-tests 2010 competitions in Singapore.

Table II The results of performance comparison between robot Adora and robot PARS II
Step field 1 Step field 2 Step field 3 Three rollers
(s) (s) (s) Ramp (s) Stairs (s) (s)
T D T D T D T D T D T D Total timea
Adora Trained operator 31.2 1.8 58.2 2.56 87 3.6 25.2 1.76 36 1.2 26.6 2.08 150.6
Beginner operator 33.2 2.6 63.2 3.44 105 7.6 28.2 3.04 44.6 3.28 36.6 3.12 169.2
PARS II Trained operator 32.4 1.7 80.8 2.24 – – 23.8 1.44 42.6 1.92 – – 179.6
Beginner operator 39.2 4.2 95.2 5.04 – – 29.6 3.52 55.6 4.08 – – 219.6
Note: aSince PARS II did not pass step field 3 and three-rollers missions, the total time does not include these missions’ times for Adora

484
 Tele-operative rescue robot with a novel track arrangement Industrial Robot: An International Journal
Nima Enayati and Farid Najafi Volume 38 · Number 5 · 2011 · 476 –485

At the time, we are working on designing a 3-DOF Koyanagi, E., Ooba, Y., Yoshida, S. and Hayashibara, Y.
manipulator arm with a gripper to add the following (2004), “Toin Pelican”, available at: [Link]
abilities to the robot: manipulating objects and elevating gov/2004_competitions (accessed 20 November 2010).
sensors to a height of 1.4 m for victim searcht. Future work Liu, J., Wang, Y., Ma, S. and Li, B. (2005), “Analysis of
focuses on development of a haptic device to control the stairs-climbing ability for a tracked reconfigurable modular
manipulator, adding vision processing for victim detection, robot”, Proceedings of IEEE International Workshop on Safety,
and development of the map generation method used on our Security and Rescue Robotics, Kobe, Japan, June, pp. 36-41.
autonomous robot and implementing it on Adora. Masayuki, A., Takayama, T. and Hirose, S. (2004),
Furthermore, we are considering adding some autonomous “Development of ‘Souryu-III’: connected crawler vehicle
abilities like automatic stair detection and climbing. for inspection inside narrow and winding spaces”,
Proceedings of the IEEE/RSJ International Conference on
Summary Intelligent Robots and Systems (IROS 2004 ), Sendai, Japan,
Vol. 1, pp. 52-7.
This paper has provided an overview of the newly developed Matthies, L., Xiong, Y., Hogg, R., Zhu, D., Rankin, A. and
Adora rescue robot. Some aspects of the system have been Kennedy, B. (2000), A Portable, Autonomous, Urban
discussed: mechanical, electrical, computer, software and Reconnaissance Robot, Jet Propulsion Laboratory,
initial tests. We have tested this robot on a standard test arena California Institute of Technology, Pasadena, CA,
and discussed how this system demonstrates improvements pp. 91-109.
over its predecessors. Adora is a tele-operated rescue robot Michaud, S., Schneider, A., Bertrand, R., Lamon, P.,
with considerable capabilities. By applying some minor Siegwart, R., Winnendael, M.V. and Schiele, A. (2002),
modifications, the introduced platform can also be “SOLERO: solar-powered exploration Rover”, Proceedings
implemented in different fields like firefighting operations,
Downloaded by Universite Laval At 07:06 01 May 2016 (PT)

of the 7th ESA Workshop on Advanced Space Technologies

anti-terrorist and street battle and mining.
for Robotics and Automation ( ASTRA2002 ), Noordwijk,
The Netherlands.
References Moosavian, S.A., Kalantari, A., Semsarilar, H., Aboosaeedan, E.
Bekker, M.G. (1969), Introduction to Terrain-vehicle Systems, and Mihankhah, E. (2009), “ResQuake: a tele-operative
The University of Michigan, Ann Arbour, MI. rescue robot”, ASME Journal of Mechanical Design, July.
Cavallin, K. and Svensson, P. (2009), “Semi-autonomous, Sheh, R. (2006), “The building of redback – a low-cost
tele-operated search and rescue robot”, MS thesis, advanced mobility platform”, available at: [Link]
Department of Computing Science, Umea University, [Link]/aboutme/[Link] (accessed
Umea. 19 November 2010).
Hirose, S. (1993), Biologically Inspired Robots (Snake-like Shigley, J.E. and Mischke, C.R. (2003), Mechanical
Locomotor and Manipulator ), Oxford University Press, Engineering Design, 6th ed., McGraw-Hill, New York, NY.
Oxford. Shiroma, N., Chiu, Y., Kamegawa, T. and Matsuno, F.
Kadous, M.W., Sheh, R. and Sammut, C. (2005), “Caster: (2004), “Development of rescue robotic systems for both
a robot for urban search and resuce”, Technical Report, daily and emergency use”, paper presented at the 1st IEEE
University of New South Wales, Sydney. Technical Exhibition Based Conference on Robotics and
Kamegawa, T., Yamasaki, T. and Matsuno, F. (2005), Automation (TExCRA2004), Tokyo.
“Evaluation of snake-like rescue robot ‘KOHGA’ for
usability of remote control”, Proceedings of the 2005 IEEE
Corresponding author
International Workshop on Safety, Security and Rescue
Robotics, Kobe, Japan, June. Nima Enayati can be contacted at: [Link]@[Link]

To purchase reprints of this article please e-mail: reprints@[Link]

Or visit our web site for further details: [Link]/reprints

485