Development of a Remote Handling Robot for the Maintenance of ...

Research ArticleDevelopment of a Remote Handling Robot forthe Maintenance of an ITER-Like D-Shaped Vessel

Peihua Chen and Qixin Cao

State Key Laboratory of Mechanical System and Vibration Research Institute of RoboticsShanghai Jiao Tong University Shanghai 200240 China

Correspondence should be addressed to Peihua Chen cphsjtueducn

Received 22 January 2014 Revised 20 May 2014 Accepted 20 May 2014 Published 27 August 2014

Academic Editor Arkady Serikov

Copyright copy 2014 P Chen and Q CaoThis is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Robotic operation is one of themajor challenges in the remotemaintenance of ITER vacuum vessel (VV) and future fusion reactorsas inner operations of Tokamak have to be done by robots due to the internal adverse conditions This paper introduces a novelremote handling robot (RHR) for themaintenance of ITER-likeD-shaped vesselThemodular designedRHRwhich is an importantpart of the remote handling system for ITER consists of three parts an omnidirectional transfer vehicle (OTV) a planar articulatedarm (PAA) and an articulated teleoperated manipulator (ATM) The task of RHR is to carry processing tools such as the viewingsystem leakage detector and electric screwdriver to inspect andmaintain the components installed inside the D-shaped vesselThekinematics of the OTV as well as the kinematic analyses of the PAA and ATM is studied in this paper Because of its special lengthand heavy payload the dynamics of the PAA is also investigated through a dynamic simulation system based on robot technologymiddleware (RTM) The results of the path planning workspace simulations and dynamic simulation indicate that the RHR hasgood mobility together with satisfying kinematic and dynamic performances and can well accomplish its maintenance tasks in theITER-like D-shaped vessel

1 Introduction

The concept of magnetic confinement fusion (MCF) haslong been explored to address the feasibility of nuclearfusion energy The InternationalThermonuclear Experimen-tal Reactor (ITER) project is an international collaborationto investigate the scientific and technological feasibility ofMCF which may be operated to test the future green energyroadmap as well as related facilities and technologies for theworld before the middle of the 21st century [1ndash3] Due tothe internal adverse radiation conditions inner operationsof Tokamak have to be done by robot manipulators Thusrobotic operation is one of the major challenges for ITER andfuture fusion reactors [4] Owing to the plasma interactionssome in-vessel components are expected to erode enoughto be repaired or replaced several times during the lifetimeof the Tokamak In particular during the installation orafter the shutdown it is important and necessary for theremote handling (RH) system to execute the inspectionand maintenance tasks in the vacuum vessel (VV) The RH

system is a key component in ITER operation both forscheduled maintenance and unexpected situations [5] Themaintenance ofVV involves various tasks such as transporta-tion inspection dust removal testing and other intelligentmaintenance Whilst ensuring confinement of radioactivedust contamination inside theVV the robotmust be operatedremotely because it has no radiation shielding capabilitiesDesign efforts have been carried on at CEA LIST to develop along reach articulated arm for ITER-like environment whichincludes the Porteur Articule en Cellule (PAC) manipulator[6] and the articulated inspection arm (AIA) [7] The PACis a long reach arm with 11 degrees of freedom (DOF) andless than 30 kg weight developed at CEA LIST mainly forAREVA-NC applications The AIA is an 8m long multilinkcarrier with five modules of 160mm diameter dedicated toperform inspection tasks on Tore Supra but without othermaintenance functionality Keller et al [8] used a CEA LISTVirtual Reality simulation software to study the accessibilityof the working zone located in front of the Bioshield andthe feasibility of connectingdisconnecting pipes using a set

Hindawi Publishing CorporationScience and Technology of Nuclear InstallationsVolume 2014 Article ID 526520 9 pageshttpdxdoiorg1011552014526520

2 Science and Technology of Nuclear Installations

(a) (b)

Figure 1 ITER Tokamak [26] and the small-scale Tokamak mockup

LGRAxis 1

Axis 2

Axis 3

Axis 4Axis 5

Axis 6

Axis 7

Axis 8Axis 9

Axis 10

Axis 11Axis 12Axis 13

ATM

OTV

Omniwheel

PAA

XY

Z

(a) (b)

Figure 2 The CAD view of RHR (a) and the actual RHR (b)

of standard industrial arms However the industrial robotthey employed was not always suitable and flexible in theITER environment for its large size and huge controllerbox As intelligent maintenance requires high accuracy anda large workspace commercially available industrial robotscannot be directly used To this end a remote handling robot(RHR) was developed by the Research Institute of Roboticsin Shanghai Jiao Tong University (RIRSJTU) China tostudy the intelligent maintenance and remote teleoperationof ITER-like D-shaped vessel as an MCF facility Meanwhilea small-scale (110) Tokamak mockup (Figure 1) was alsodesigned and set up to test the robot performance and thefeasibility of the remote handling system In this project theaim of this work is to introduce the new remote handlingrobot analyze its kinematic and dynamic performances anddiscuss the feasibility of the mechanical design and controlmethod verified at least in the simulation

The rest of this paper is organized as follows In Section 2the novel RHR is introduced and the important study issues

of the robot are also presented The kinematic and dynamicanalysis of the RHR is presented in Section 3 Section 4 isdevoted to illustrative simulations and the correspondingresults In Section 5 conclusions are drawn and the futurework is discussed

2 The Remote Handling Robot (RHR)

The RHR as shown in Figure 2 is a mobile robot with amultiple degrees of freedom articulated arm The RHR iscomposed of an omnidirectional transfer vehicle (OTV) aplanar articulated arm (PAA) and an articulated teleoperatedmanipulator (ATM)

The OTV has four 810158401015840 heavy-duty Mecanum wheels anda linear guide rail (LGR) with a long travel of 2100mm andmaximum feed rate of 01ms Although the AIA robot alsohas a transfer vehicle it has a limited mobility and has to beoperated by winches and guide rails on the ground [9] Inthis paper with the special mechanism of omnidirectional

Science and Technology of Nuclear Installations 3

wheels the OTV is able to perform 3-DOF motion onthe two-dimensional plane which achieves translationaland rotational movements simultaneously along arbitrarydirection in the Tokamak room The maximum linearvelocity of the OTV is 03ms The OTV together with LGRweighs about 226 kg and supports 200 kg loads The LGRcould make the PAA and ATM be deployed into the D-shaped vessel through the ball screw nut fittingThe Beckhoffhigh-performance AC servomotors are adopted as theactuators in the OTV because they are available in stainlesssteel designs and characterized by high dynamics energyefficiency low costs and real-time (as they adopted theEtherCAT solutions) Connected to the OTV via the LGRthe PAA is a 2m longmultilink armwith amaximumpayloadup to 18 kg The PAA has five modular planar joints witheach module being 350mm in length and the last link being250mm in length which give to the PAA six DOFs in total asdifferent colors indicated in Figure 2(a) Each module of thePAAmainly includes a Maxon DC servomotor 2 bevel gearsincremental and absolute encoders a harmonic drive gearand EtherCAT bus driver The bevel gears can transform therotational direction of the motorrsquos axis and make the outputshaft of the joint vertical The incremental and absoluteencoders are used to measure the relative position of theservomotor and the absolute position of the output shaft ofthe joint respectively The harmonic drive gear is adoptedhere because it has no backlash compactness and lightweight and high gear ratio which are very important factorsfor a long reach multiple-link arm The EtherCAT bus driveris employed to perform the real-time communication andcontrol All the elements of themodule are installed inside thelink and aremostly made of aluminum alloy or stainless steelThe ATM installed on the end link of the PAA is a six-DOFarticulated manipulator which weighs approximately 13 kgwith a load capacity of 15 kgThe structure of the ATM is likean industrial robot which also has a spherical wrist whereinthe three wrist axes (axes 11ndash13 as shown in Figure 2(a))intersect at a single point but is redesigned to adapt to theITER-like D-shaped vessel environment All the joints of theATM are also driven by actuators with harmonic drive gearsThe ATM is expected to perform the poloidal inspectionand maintenance process when it is at the right toroidalposition in the D-shaped vessel With large workspace andadequate flexibility the PAA and ATM can be deployed intothe D-shaped vessel for the inspection or maintenance taskthrough the equatorial access port The PAA the ATM andthe LGRmake up the arm systemThis paper investigates theimportant and necessary issues for the study of the RHR pathplanning of the OTV kinematic and dynamic analyses of thePAA and kinematic and reachability analyses of the ATM

As a result of the modular design of the whole robot theoperation of RHR is classified into three steps (1) movingthe OTV to the target equatorial access port and performinga docking with the port (2) deploying the PAA and ATMintoout of the D-shaped vessel with combined motions ofaxes 1ndash7 in the horizontal plane and (3) spreading out theATM to execute the maintenance pertinent task The designprocedures and analysis methodologies may be adopted or

1

23

4

1205963 1205962

1205964 1205961l

X

Y

O

120596

120573

Figure 3 Four-wheeled OTV

developed further by other researchers for the developmentof relevant remote handling robots for ITER

3 Robot Kinematics and Dynamics

31 Kinematics of the OTV To perform path planning ofthe OTV the kinematics of the vehicle should first besolved Figure 3 shows the coordinate system of the OTV andthe configuration of omnidirectional wheels it adopts Theconstraint equation between the angular velocity of the wheeland the velocity of the OTV (

0 1199100 120596)119879 is represented by

(1205961 1205962 1205963 1205964)119879

=119881

119882T sdot(0 1199100 120596)119879

(1)

where 119881119882T isin R4times3 denotes the transformation matrix

from the velocity of the vehicle to that of the four wheelsIt is possible to obtain the equations that determine theOTVrsquos speed related with wheelsrsquo speeds but the matrix119881

119882T associated with (1) is not square This is because the

four-wheeled system is redundant In order to obtain theomnidirectional characteristic of theOTV thematrix 119881

119882Thas

to be full column rank that is rank ( 119881119882T) = 3 As for theOTV

the transformation matrix 119881119882T can be derived as

119881

119882T =

1

119903

[[[

[

1 1 119897 (cos120573 + sin120573)

minus1 1 119897 (cos120573 + sin120573)

minus1 minus1 119897 (cos120573 + sin120573)

1 minus1 119897 (cos120573 + sin120573)

]]]

]

(2)

where 119903 denotes the radius of the omniwheel 119897 denotes thedistance between the center of the omniwheel and the centerof the OTV and 120573 denotes the angle between 119897 and the119884 axisas depicted in Figure 3

32 Kinematic Analysis of the Arm System Owing to themodular design and operation procedure of RHR as men-tioned in Section 2 the kinematics of the PAA and ATMcan be calculated separately in the arm system As theLGR has only one DOF and works in the procedure ofthe PAArsquos deployment into the D-shaped vessel it will beincluded in the kinematic analysis of the PAA acting aslink 1 According to D-H convention and transformation


theory between neighboring link coordinate systems thehomogeneous transformationmatrix of the coordinate frame119894 as represented in the coordinate frame 119894minus1 can be expressedas [10]

119894minus1

119894T =

[[[

[

119888120579119894minus119904120579119894119888120572119894

119904120579119894119904120572119894

119886119894119888120579119894

119904120579119894

119888120579119894119888120572119894

minus119888120579119894119904120572119894119886119894119904120579119894

0 119904120572119894

119888120572119894

119889119894

0 0 0 1

]]]

]

(119894 = 1 2 sdot sdot sdot 13)

(3)

where the four quantities 120579119894 119886119894 119889119894 and 120572

119894are parameters

associated with link 119894 and joint 119894 119888 equiv cos and 119904 equiv sin Asillustrated in Section 2 once the PAA is introduced into D-shaped vessel in the toroidal direction the axes 1ndash7 (Figure 2)will be fixed The overall forward kinematics 0

13T for the arm

system can be divided into two kinematic equations 07T and

7

13T thus we have

0

13T =0

7T sdot7

13T

0

7T =0

1T sdot1

7T =0

1T 12T 23T 34T 45T 56T 67T

7

13T =7

8T 89T 910T 1011T 1112T 1213T

(4)

Consequently the forward kinematics can be solved from(4) with D-H parameters The Jacobian matrices relating thejoint velocities to the linear and angular velocities of the end-effector can be obtained using the differential transformationmethod [10] namely J

119860119871isin R6times7 for the PAA (with LGR) and

J119872

isin R6times6 for the ATMInverse kinematics is required to determine the angles

of the joints and displacement of the LGR while the poseof the end-effector of the ATM is given and is used in theposition control of the arm system Given a target pose p =

(119909 119910 119911 120572 120573 120574)119879 the inverse kinematics can be divided into

two steps First the end-link pose of the PAA is derived asp119860

= (119909 1199101015840 10345 0 0 120574)119879 where the height of the PAA is

10345mm and 1199101015840 can be obtained from the equation of thecenter circle of the D-shaped vessel which is (119909 minus 247196)

2+

1199102 = 5602The joint velocities of the PAA (with LGR) in jointspace can be derived as

q119860119871

= J+119860119871p119860 (5)

where q119860119871

= (1199021 1199022 1199023 1199024 1199025 1199026 1199027)119879 is the vector of joint

angles of the PAA (with LGR) and J+119860119871

is the pseudoinverseof the Jacobian matrix J

119860119871 The resolved motion rate control

(RMRC) method [11] which is a kind of numerical solutionsis then chosen for the integration of the joint velocities from0 to 119905119891to produce the joint angles

q119860119871

(119905119891) = q1198601198710

+ int119905119891

0

q119860119871

(119905) 119889119905 (6)

where q119860119871

(119905119891) = q

119860119871 q1198601198710

is the vector of initial jointangles of the PAA (with LGR) and q

119860119871(119905) is derived from

(5) Second as we could obtain 013T and 0

7T from p and p

119860

respectively 713T could be obtained from (4) and the target

pose in the ATM coordinate system could be derived as p119872

Then the joint velocities of the ATM could be derived from

q119872

= J+119872p119872 (7)

where q119872

= (1199028 1199029 11990210 11990211 11990212 11990213)119879 is the vector of joint

angles of theATMand J+119872is the pseudoinverse of the Jacobian

matrix J119872 The joint angles of the ATM can also be obtained

by utilization of the RMRC method

33 Dynamics of the PAA Aside from the kinematics of thearm system dynamics is another important aspect for a robotespecially for the PAA in this RHR because of the speciallength and heavy payload of the arm The dynamics also hasa significant impact on the development of the mechanicaldesign and control system Excellent dynamic performancecontributes to a simple and light structure as well as a simpleand reliable control system Following [12] in the absenceof friction and disturbance the dynamics of the PAA can bedescribed as

M (q) q + C (q q) q + g (q) = 120591 (8)

where q isin R6 is the vector of the generalized joint coordi-nates 120591 isin R6 is the vector of the generalized torques actingon the PAA M(q) isin R6times6 is the inertia matrix C(q q)q isin

R6 is the vector of Corioliscentrifugal forces and g(q) =

120597H(q)120597q isin R6 is the gradient of the potential functionH(q)However the computational cost of (8) is enormous

[10] 119874(1198994) which made it infeasible for real-time use in

actual application So the recursive Newton-Euler algorithm(RNEA) with 119874(119899) computational complexity [13] is usedto compute the inverse dynamics of the PAA The RNEAstarts at the base and works outward adding the velocity andacceleration of each joint in order to determine the velocityand acceleration of each link Then working from the end-link back to the base it computes the forces and momentsacting on each link and thus the joint torquesThe RNEA canbe written in the form

120591 = 119868119863 (q q q) (9)

During the motion control of the PAA there will beuncertainty in the inertia parameters unmodeled forces orexternal disturbances Thus a feedback term is needed tocompensate for any errors due to the uncertainty of therobot dynamics Here the inverse dynamics control method[14 15] is used to compute the joint torques and deal with thedynamic uncertainty wherein a PD controller is adopted asthe feedback termThe inverse dynamics control method canbe given by

120591 = 119868119863 (q q q119889+ K119863e + K119875e) (10)

where e = q119889minusq is the vector of joint position errorsK

119875and

K119863are the position and velocity gain matrices respectively

and q = q119889+ K119863e + K

119875e The control law of (10) which

contains the terms K119875e +K119863e that are actually premultiplied

by the inertial matrix is not a linear controller as the PD


Start

positionGoal

position

(a)

Original dataPolynomial fitting pathOrientation of OTV

Y(m

)

X (m)

0

0

1

minus1

minus1

minus2

minus2

minus3

minus3

minus4

minus4

minus5

minus5

minus6

minus6

minus7

minus7minus8minus8

(b)

Figure 4 Planning result of RRT-Connect planner (a) and the polynomial fitting path with respect to the orientation of the vehicle (b)

controller The inverse dynamics control is one of the model-based motion control approaches for manipulators In ourprevious work [15] we have discussed the details about theinverse dynamics control its application on the trajectorycontrol of the articulated robot and comparisons with PDcontroller In practice the joint space trajectory is alwaysplanned first that is the desired trajectory of motion q

119889(119905)

and its derivatives q119889(119905) and q

119889(119905) could be obtained Thus

we usually use 119868119863(q119889 q119889 q119889+ K119863e + K

119875e) to compute the

required joint torques to achieve the desired motions of therobot [16] The linear error equation is then derived as

e + K119863e + K119875e = 0 (11)

If the gain matrices K119875and K

119863are chosen as symmetric

positive-definite matrices then the closed-loop system islinear decoupled and globally stable

4 Experimental Results

41 Path Planning of the OTV Since the environment ofTokamak building is comprised of static and well-structuredscenarios the path planning of the OTV could be executedoffline and the planned paths may be stored in a teaching filefor further use Fonte et al [17] proposed a motion planningmethodology for the transfer cask system (TCS) to operatein the cluttered and static ITER environment Recently theyproposed the cask andplug remote handling system (CPRHS)to provide the means for the remote transfer of invesselcomponents and remote handling equipment between thehot cell building (HCB) and the Tokamak building in ITERalong predefined optimized trajectories [18ndash20] However

they used a rhombic kinematic configuration with two pairsof drivable and steerable wheels in the vehicle which hasa limited mobility comparing to the OTV adopted in thispaper In addition challenges about how to evaluate the celldecompositions and how to improve the efficiency of thepath planning still exist In this paper we employed the openmotion planning library (OMPL) [21] and the RRT-Connect[22] algorithm to perform the path planning between astart position and the target position for docking to theequatorial access port as presented in Figure 4(a) The pathplanning methodology takes the geometry of the OTV intoconsideration which will shorten the planning time Theaverage runtime of the path planning is approximately 026s on a machine with an Intel i5-2400 CPU (31 Ghz) and 4GBRAM (the same hereinafter) However the initial path doesnot guarantee the path smoothness Thus the polynomialfitting method is utilized to fit the discrete trajectory dataand obtain a smooth and continuous path for the robotThe polynomial fitting path is shown in Figure 4(b) whereinthe green arrows denote the OTV orientation After theOTV moves to the target position which is always a littledistance away from the equatorial access port it will executethe docking process as introduced in Section 2 During thisprocess the control room will teleoperate the OTV remotelyvia EtherCAT bus tomove towards the equatorial access portwherein the orientation of the OTV could also be adjustedremotely and conveniently In order to realize the docking tothe equatorial access port of the ITER-like D-shaped vesselwe designed a pair of mounting flanges on the OTV and theaccess port respectively In the mounting flange on the OTVside there are two taper holes for guiding and aligning to thecircular truncated cone fixed on the equatorial access port


Figure 5 Schematic view of RHR integration on the Tokamakmockup (top view)

42 Workspace Simulation of the PAA A large workspaceserves an important function for the PAA during the main-tenance work in the D-shaped vessel Concretely the PAAcould reach asmuch area as possible in the horizontal plane ofthe D-shaped vessel from the same access port efficiency willbe improved and the number of robots needed to fulfill themaintenance of ITER-like D-shaped vessel will be reducedPeng et al [23] introduced a long reach multiarticulatedmanipulator with ten degrees of freedom (DOFs) formed bya main robot and an end-effector However this robot couldnot cover all of the horizontal space in the EASTVV from thesame access port In this paper the PAA of RHR could reachalmost every corner of the horizontal plane of the D-shapedvessel with the associated movements of the LGR and PAAas depicted in Figure 5

The workspace of the PAA has been investigated byMATLAB simulation utilizing Monte Carlo method as arandom sampling numerical approach to create thousands ofrandom joint series in the joint spaceThe forward kinematicmodel is then used to transform the joint space to theCartesian space of the PAArsquos end-effector which determinesthe workspace of the PAA Figure 6 shows the workspace ofthe PAA (displacement of the LGR is 1500mm) within the D-shaped vessel in the horizontal plane Different colors denotedifferent density of the reachable workspace at that positionwherein the amount of possible reachable configurationbecomes larger when the color changes from green to red

According to the definition of kinematic redundancy ofa robot in literature [24] the PAA together with the LGR iskinematically redundant as the dimension of the joint space isgreater than the dimension of end-effector space During thewhole manipulatorrsquos insertion into the D-shaped vessel thePAA has to be deployed into the equatorial access port firstvia motion planning In this process given a target postureinverse kinematics of the PAA (with LGR) is employed totransform from task space to joint space When the goalconfiguration in the joint space is obtained and valid themotion planning based on RRT-Connect [22] is adopted tofind a collision-free path from the start configuration to thegoal configuration

43 Workspace Simulation of the ATM TheATM is expectedto perform the maintenance work inside the D-shaped vesselafter being brought to a right toroidal position by the PAAOwing to the planar feature of the PAA the ATM needs toreach as much space as possible with its own mechanical

0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1

Excircle of D-shaped vessel

Inner circle of D-shaped vessel

Center circle of D-shaped vessel

minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)

Figure 6 Workspace of the PAA in the horizontal plane inside theD-shaped vessel

1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)

Figure 7Workspace simulation of the ATM in theD-shaped vessel

structure so that the maintenance work can be done Consid-ering the workspace requirement and the size of a D-shapedcross section of the vessel we designed a six-DOF ATMwithoptimized link lengths and mass distribution (see Figure 2)As the PAA can reach almost every corner in the horizontalplane of theD-shaped vessel only the reachability of theATMat one position along the toroidal direction has to be analyzedfor example the position on the center circle of D-shapedvessel Figure 7 depicts that the PAA deploys the ATM to theposition of (293196 0 9595) where the pedestal height of theATM is 75mm lower than that of the PAAThe workspace ofthe ATM (green color in the figure) is simulated with respectto the cross section of the 3D D-shaped vessel model (bluecolor in the figure) in MATLAB The result shows a perfectcoverage of the inner surface of the vessel in the D-shapedcross section

44 Dynamic Simulation of the PAA To investigate thedynamic properties of the PAA a dynamic simulation systembased on robot technology middleware (RTM) [25] wasdeveloped to study the dynamic performance for the purposeof verifying the current designrsquos reliability and investigating


Inversedynamics RobotMotion

planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q

Figure 8 Dynamic simulation system block scheme

the control algorithmrsquos feasibility as depicted in Figure 8In the dynamic simulation we used the inverse dynamicscontrol method which is based on the RNEA to calculate thejoint torques that have to be applied on the PAArsquos joints Theimplementation of (10) is realized in the controller moduleof the dynamic simulation system as shown in Figure 8 Inorder to make use of the planned data created by the motionplanner to obtain the computed torques for controlling themotion of the robot we make the inputs of the inversedynamics controller as q

119889 q119889 and q

119889+K119863e+K119875e to compute

the required joint torques The PAA is required to follow thepath from the initial position (Figure 2) to the fully unfoldedposition (Figure 5) with or without 15 kg payload (include themass of the ATM) The LGR is fixed during the simulationThe controller gains are chosen as K

119875= 10000I and K

119863=

120I The motion planning for the path is done with jointspace planning and Cartesian space planning in the motionplanner module From the simulation the joint coordinatesvelocities accelerations and torques along the path can beobtained The desired joint trajectories are chosen to befifth-order polynomials and are shown in Figure 9 The jointvelocity evolution of the PAA is presented in Figure 10 Theresults reveal that themotions velocities and accelerations ofeach joint evolved smoothly Figure 11 shows a comparison oftorque evolution under no payload with that under a payloadof 15 kg The torque value of each joint is in the limitation ofcorresponding selected actuator The value is larger when thePAA has a payload

5 Conclusions and Future Work

In this paper the kinematic and dynamic analyses of anovel RHR have been studied for the remote maintenanceof an ITER-like D-shaped vessel The RHR consists of threeparts namely the OTV the PAA and the ATM The PAA isconnected to the OTV through the LGR whereas the ATMis installed on the end link of the PAA The path planningand trajectory optimization of the OTV show that the RHRhas good mobility in the vehicle part The modular design ofthe robot facilitates simple development and easy integrationand offers the RHR simple kinematic formulations Thekinematic analysis of the arm system and the workspacesimulations of the PAA and ATM show that the RHR issufficiently dexterous to reach the needed workspace throughthe combinedmotions of the PAA and ATM such that at leastone RHR is required to fulfill the inspection or maintenancetasks of thewhole invessel surfacesWorkspace simulations ofthe PAA and ATM together with the dynamic simulation of

0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)

Desired joint trajectories

minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6

Figure 9 Desired joint trajectories for the dynamic simulation

0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)

Joint velocity versus time

minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6

Figure 10 Joint velocity evolution during the dynamic simulation

the PAAusing the inverse dynamics control were studied andintegrated into the design process to verify the mechanicaldesign and feasibility of control algorithms All simulationsindicated that the RHR has superior kinematic and dynamicperformances and proved the feasibility and reliability of therobot in terms of the maintenance of the ITER-like D-shapedvessel

Considerable design analysis and investigations arerequired tomove from the present status (mechanical designsupported by simulation analysis) to the final design andapplication on the actual remote handling system Robust-ness reliability and flexibility of the RHR will be tested and


0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm

)Joint torque versus time without payload

minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)

Joint torque versus time with 15kg payload

minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)

Figure 11 Comparison of torque evolution under no payload (a) and with 15 kg payload (b)

improved during successive operating campaigns in a routinemaintenance mode in the future

Conflict of Interests

The authors declare that there is no conflict of interestsregarding the publication of this paper

Acknowledgments

This work is supported by the Major State Basic ResearchDevelopment Program of China under Grant no2011GB113005 and in part by the National High TechnologyResearch and Development Program of China under Grantno 2012AA041403 The authors would also like to thankthe anonymous reviewers and the editor for their helpfulcomments on the paper

References

[1] D Clery ldquoITERrsquos $12 billion gamblerdquo Science vol 314 no 5797pp 238ndash242 2006

[2] A E Costley ldquoTowards diagnostics for a fusion reactorrdquo IEEETransactions on Plasma Science vol 38 no 10 pp 2934ndash29432010

[3] G Federici R Kemp D Ward et al ldquoOverview of EU DEMOdesign and RD activitiesrdquo Fusion Engineering and Design 2014

[4] L Gargiulo P Bayetti V Bruno et al ldquoOperation of an ITERrelevant inspection robot on Tore Supra tokamakrdquo FusionEngineering and Design vol 84 no 2ndash6 pp 220ndash223 2009

[5] I Ribeiro C Damiani A Tesini S Kakudate M Siuko andC Neri ldquoThe remote handling systems for ITERrdquo FusionEngineering and Design vol 86 no 6ndash8 pp 471ndash477 2011

[6] J Chaifoun C Bidard D Keller Y Perrot and G PiolainldquoDesign and flexiblemodeling of a long reach articulated carrierfor inspectionrdquo in Proceedings of the IEEERSJ InternationalConference on Intelligent Robots and Systems (IROS rsquo07) pp4013ndash4019 November 2007

[7] L Gargiulo P Bayetti V Bruno et al ldquoDevelopment of an ITERrelevant inspection robotrdquo Fusion Engineering and Design vol83 no 10-12 pp 1833ndash1836 2008

[8] D Keller C Dechelle L Doceul et al ldquoITER EquatorialPort plug engineering design and remote handling activitiessupported by Virtual Reality toolsrdquo Fusion Engineering andDesign vol 86 no 9ndash11 pp 2105ndash2108 2011

[9] D Keller J-P Friconneau and Y Perrot ldquoAn ITER relevantrobot for remote handling on the road to operation on ToreSuprardquo in Advances in Service Robotics H S Ahn Ed InTechRijeka Croatia 2008

[10] M W Spong S Hutchinson and M Vidyasagar Robot Model-ing and Control JohnWiley amp Sons New York NY USA 2006

[11] D E Whitney ldquoResolved motion rate control of manipulatorsand human prosthesesrdquo IEEE Transactions on Man-MachineSystems vol 10 no 2 pp 47ndash53 1969

[12] J J Craig Introduction to Robotics Mechanics and ControlPearson Education New York NY USA 3rd edition 2005

[13] R Featherstone Rigid Body Dynamics Algorithms SpringerNew York NY USA 2008

[14] M W Spong and R Ortega ldquoOn adaptive inverse dynamicscontrol of rigid robotsrdquo IEEE Transactions on Automatic Con-trol vol 35 no 1 pp 92ndash95 1990

[15] P Chen and Q Cao ldquoTrajectory control of the articulated robotbased on inverse dynamicsrdquo Journal of Huazhong Universityof Science and Technology (Natural Science Edition) vol 41supplement I pp 17ndash20 2013

[16] P I Corke ldquoA robotics toolbox for MATLABrdquo IEEE Roboticsand Automation Magazine vol 3 no 1 pp 24ndash32 1996


[17] D Fonte F Valente A Vale and I Ribeiro ldquoA motion plan-ning methodology for rhombic-like vehicles for ITER RemoteHandling operationsrdquo in Proceedings of the 7th IFAC Symposiumon Intelligent Autonomous Vehicles (IAV rsquo10) pp 443ndash448September 2010

[18] A Vale D Fonte F Valente J Ferreira I Ribeiro and CGonzalez ldquoFlexible path optimization for the cask and plugremote handling system in ITERrdquo Fusion Engineering andDesign vol 88 no 9-10 pp 1900ndash1903 2013

[19] R Ventura J Ferreira I Filip and A Vale ldquoLogistics man-agement for storing multiple cask plug and remote handlingsystems in ITERrdquo Fusion Engineering and Design vol 88 no9-10 pp 2062ndash2066 2013

[20] D Locke C Gonzalez Gutierrez C Damiani J-P Friconneauand J-PMartins ldquoProgress in the conceptual design of the ITERcask and plug remote handling systemrdquo Fusion Engineering andDesign 2014

[21] I A Sucan M Moll and L Kavraki ldquoThe open motionplanning libraryrdquo IEEE Robotics and AutomationMagazine vol19 no 4 pp 72ndash82 2012

[22] J J Kuffner Jr and S M la Valle ldquoRRT-connect an efficientapproach to single-query path planningrdquo in Proceedings of theIEEE International Conference on Robotics and Automation pp995ndash1001 April 2000

[23] X B Peng Y T Song C C Li M Z Lei andG Li ldquoConceptualdesign of EAST flexible in-vessel inspection systemrdquo FusionEngineering and Design vol 85 no 7ndash9 pp 1362ndash1365 2010

[24] E S Conkur and R Buckingham ldquoClarifying the definition ofredundancy as used in roboticsrdquoRobotica vol 15 no 5 pp 583ndash586 1997

[25] N Ando T Suehiro K Kitagaki T Kotoku andW Yoon ldquoRT-Middleware distributed component middleware for RT (RobotTechnology)rdquo in Proceedings of the IEEE IRSRSJ InternationalConference on Intelligent Robots and Systems (IROS rsquo05) pp3555ndash3560 August 2005

[26] ITER ITERmdashThe Way to New Energy 2013 httpswwwiterorg

TribologyAdvances in

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

FuelsJournal of


Journal ofPetroleum Engineering


Industrial EngineeringJournal of


Power ElectronicsHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Advances in

CombustionJournal of


Journal of


Renewable Energy

Submit your manuscripts athttpwwwhindawicom


StructuresJournal of


RotatingMachinery


EnergyJournal of


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal ofPhotoenergy


Nuclear InstallationsScience and Technology of


Solar EnergyJournal of


Wind EnergyJournal of


Nuclear EnergyInternational Journal of


High Energy PhysicsAdvances in

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014


(a) (b)

Figure 1 ITER Tokamak [26] and the small-scale Tokamak mockup

LGRAxis 1

Axis 2

Axis 3

Axis 4Axis 5

Axis 6

Axis 7

Axis 8Axis 9

Axis 10

Axis 11Axis 12Axis 13

ATM

OTV

Omniwheel

PAA

XY

Z

(a) (b)

Figure 2 The CAD view of RHR (a) and the actual RHR (b)

of standard industrial arms However the industrial robotthey employed was not always suitable and flexible in theITER environment for its large size and huge controllerbox As intelligent maintenance requires high accuracy anda large workspace commercially available industrial robotscannot be directly used To this end a remote handling robot(RHR) was developed by the Research Institute of Roboticsin Shanghai Jiao Tong University (RIRSJTU) China tostudy the intelligent maintenance and remote teleoperationof ITER-like D-shaped vessel as an MCF facility Meanwhilea small-scale (110) Tokamak mockup (Figure 1) was alsodesigned and set up to test the robot performance and thefeasibility of the remote handling system In this project theaim of this work is to introduce the new remote handlingrobot analyze its kinematic and dynamic performances anddiscuss the feasibility of the mechanical design and controlmethod verified at least in the simulation

The rest of this paper is organized as follows In Section 2the novel RHR is introduced and the important study issues

of the robot are also presented The kinematic and dynamicanalysis of the RHR is presented in Section 3 Section 4 isdevoted to illustrative simulations and the correspondingresults In Section 5 conclusions are drawn and the futurework is discussed

2 The Remote Handling Robot (RHR)

The RHR as shown in Figure 2 is a mobile robot with amultiple degrees of freedom articulated arm The RHR iscomposed of an omnidirectional transfer vehicle (OTV) aplanar articulated arm (PAA) and an articulated teleoperatedmanipulator (ATM)

The OTV has four 810158401015840 heavy-duty Mecanum wheels anda linear guide rail (LGR) with a long travel of 2100mm andmaximum feed rate of 01ms Although the AIA robot alsohas a transfer vehicle it has a limited mobility and has to beoperated by winches and guide rails on the ground [9] Inthis paper with the special mechanism of omnidirectional




1

23

4

1205963 1205962

1205964 1205961l

X

Y

O

120596

120573






(1205961 1205962 1205963 1205964)119879

=119881

119882T sdot(0 1199100 120596)119879

(1)





119882Thas



119881

119882T =

1

119903

[[[

[

1 1 119897 (cos120573 + sin120573)

minus1 1 119897 (cos120573 + sin120573)


1 minus1 119897 (cos120573 + sin120573)

]]]

]

(2)





119894minus1

119894T =

[[[

[

119888120579119894minus119904120579119894119888120572119894

119904120579119894119904120572119894

119886119894119888120579119894

119904120579119894

119888120579119894119888120572119894

minus119888120579119894119904120572119894119886119894119904120579119894

0 119904120572119894

119888120572119894

119889119894

0 0 0 1

]]]

]

(119894 = 1 2 sdot sdot sdot 13)

(3)




13T for the arm


7

13T thus we have

0

13T =0

7T sdot7

13T

0

7T =0

1T sdot1

7T =0

1T 12T 23T 34T 45T 56T 67T

7

13T =7

8T 89T 910T 1011T 1112T 1213T

(4)



J119872







2+


q119860119871

= J+119860119871p119860 (5)

where q119860119871






q119860119871

(119905119891) = q1198601198710

+ int119905119891

0

q119860119871

(119905) 119889119905 (6)

where q119860119871

(119905119891) = q

119860119871 q1198601198710


119860119871(119905) is derived from


7T from p and p

119860




q119872

= J+119872p119872 (7)

where q119872






M (q) q + C (q q) q + g (q) = 120591 (8)






120591 = 119868119863 (q q q) (9)


120591 = 119868119863 (q q q119889+ K119863e + K119875e) (10)


119875and


and q = q119889+ K119863e + K





Start

positionGoal

position

(a)


Y(m

)

X (m)

0

0

1

minus1

minus1

minus2

minus2

minus3

minus3

minus4

minus4

minus5

minus5

minus6

minus6

minus7

minus7minus8minus8

(b)



119889(119905)






e + K119863e + K119875e = 0 (11)













0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1




minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)


1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)






planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of




















1

23

4

1205963 1205962

1205964 1205961l

X

Y

O

120596

120573






(1205961 1205962 1205963 1205964)119879

=119881

119882T sdot(0 1199100 120596)119879

(1)





119882Thas



119881

119882T =

1

119903

[[[

[

1 1 119897 (cos120573 + sin120573)

minus1 1 119897 (cos120573 + sin120573)


1 minus1 119897 (cos120573 + sin120573)

]]]

]

(2)





119894minus1

119894T =

[[[

[

119888120579119894minus119904120579119894119888120572119894

119904120579119894119904120572119894

119886119894119888120579119894

119904120579119894

119888120579119894119888120572119894

minus119888120579119894119904120572119894119886119894119904120579119894

0 119904120572119894

119888120572119894

119889119894

0 0 0 1

]]]

]

(119894 = 1 2 sdot sdot sdot 13)

(3)




13T for the arm


7

13T thus we have

0

13T =0

7T sdot7

13T

0

7T =0

1T sdot1

7T =0

1T 12T 23T 34T 45T 56T 67T

7

13T =7

8T 89T 910T 1011T 1112T 1213T

(4)



J119872







2+


q119860119871

= J+119860119871p119860 (5)

where q119860119871






q119860119871

(119905119891) = q1198601198710

+ int119905119891

0

q119860119871

(119905) 119889119905 (6)

where q119860119871

(119905119891) = q

119860119871 q1198601198710


119860119871(119905) is derived from


7T from p and p

119860




q119872

= J+119872p119872 (7)

where q119872






M (q) q + C (q q) q + g (q) = 120591 (8)






120591 = 119868119863 (q q q) (9)


120591 = 119868119863 (q q q119889+ K119863e + K119875e) (10)


119875and


and q = q119889+ K119863e + K





Start

positionGoal

position

(a)


Y(m

)

X (m)

0

0

1

minus1

minus1

minus2

minus2

minus3

minus3

minus4

minus4

minus5

minus5

minus6

minus6

minus7

minus7minus8minus8

(b)



119889(119905)






e + K119863e + K119875e = 0 (11)













0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1




minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)


1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)






planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of



















119894minus1

119894T =

[[[

[

119888120579119894minus119904120579119894119888120572119894

119904120579119894119904120572119894

119886119894119888120579119894

119904120579119894

119888120579119894119888120572119894

minus119888120579119894119904120572119894119886119894119904120579119894

0 119904120572119894

119888120572119894

119889119894

0 0 0 1

]]]

]

(119894 = 1 2 sdot sdot sdot 13)

(3)




13T for the arm


7

13T thus we have

0

13T =0

7T sdot7

13T

0

7T =0

1T sdot1

7T =0

1T 12T 23T 34T 45T 56T 67T

7

13T =7

8T 89T 910T 1011T 1112T 1213T

(4)



J119872







2+


q119860119871

= J+119860119871p119860 (5)

where q119860119871






q119860119871

(119905119891) = q1198601198710

+ int119905119891

0

q119860119871

(119905) 119889119905 (6)

where q119860119871

(119905119891) = q

119860119871 q1198601198710


119860119871(119905) is derived from


7T from p and p

119860




q119872

= J+119872p119872 (7)

where q119872






M (q) q + C (q q) q + g (q) = 120591 (8)






120591 = 119868119863 (q q q) (9)


120591 = 119868119863 (q q q119889+ K119863e + K119875e) (10)


119875and


and q = q119889+ K119863e + K





Start

positionGoal

position

(a)


Y(m

)

X (m)

0

0

1

minus1

minus1

minus2

minus2

minus3

minus3

minus4

minus4

minus5

minus5

minus6

minus6

minus7

minus7minus8minus8

(b)



119889(119905)






e + K119863e + K119875e = 0 (11)













0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1




minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)


1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)






planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of


















Start

positionGoal

position

(a)


Y(m

)

X (m)

0

0

1

minus1

minus1

minus2

minus2

minus3

minus3

minus4

minus4

minus5

minus5

minus6

minus6

minus7

minus7minus8minus8

(b)



119889(119905)






e + K119863e + K119875e = 0 (11)













0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1




minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)


1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)






planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of























0 500 1000 1500 2000 2500 3000

0

500

1000

1500

x

z

7

y

65

4

32

ITERPAA

1




minus500

minus500

minus1000

minus1500

Y(m

m)

X (mm)


1600 1800 2000 2200 2400 2600 2800 3000 3200 3400500

600

700

800

900

1000

1100

1200

1300

1400

1500Z

(mm

)

X (mm)






planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of



















planner

Controller

minus

minus

+

+

+++

120591

Kp

KD

qdqdqd

q

q

q



119889 q119889 and q

119889+K119863e+K119875e to compute


119875= 10000I and K

119863=




0 5 10 15 20 25 30 35 40

0

05

1

15

2

25

Time (s)

Join

t ang

le (r

ad)


minus05

minus15

minus2

minus1

Joint1Joint2Joint3

Joint4Joint5Joint6


0 5 10 15 20 25 30 35 40

0

002

004

006

008

01

012

Time (s)

Join

t velo

city

(rad

s)


minus002

minus004

minus006

minus008

Joint1Joint2Joint3

Joint4Joint5Joint6





0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of


















0 5 10 15 20 25 30 35 40

0

20

40

60

80

Time (s)

Join

t tor

que (

Nm


minus20

minus40

minus60

minus80

Joint1Joint2Joint3

Joint4Joint5Joint6

(a)

0 5 10 15 20 25 30 35 40

0

20

40

60

80

100

120

Time (s)

Join

t tor

que (

Nm

)


minus80

minus60

minus40

minus20

Joint1Joint2Joint3

Joint4Joint5Joint6

(b)





Acknowledgments


References
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of
































FuelsJournal of







Advances in



Journal of


Renewable Energy





RotatingMachinery


EnergyJournal of

















Development of a Remote Handling Robot for the Maintenance of ...

Documents

Transcript of Development of a Remote Handling Robot for the Maintenance of ...