Industrial Robot: An International Journal

Cobots
Michael PeshkinJ. Edward Colgate
Article information:
To cite this document:
Michael PeshkinJ. Edward Colgate, (1999),"Cobots", Industrial Robot: An International Journal, Vol. 26 Iss 5 pp. 335 - 341
Permanent link to this document:
http://dx.doi.org/10.1108/01439919910283722
Downloaded on: 05 November 2014, At: 10:58 (PT)
References: this document contains references to 7 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 261 times since 2006*
Users who downloaded this article also downloaded:
Anna Kochan, (2004),"CLAWAR robots progress slowly but surely", Industrial Robot: An International Journal, Vol. 31 Iss 1
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

pp. 22-25
Jeanne Dietsch, Brett Kennedy, Avi Okon, Hrand Aghazarian, Mircea Badescu, Xiaoqi Bao, Yoseph Bar#Cohen, Zensheu
Chang, Borna E. Dabiri, Mike Garrett, Lee Magnone, Stewart Sherrit, (2006),"Lemur IIb: a robotic system for steep terrain
access", Industrial Robot: An International Journal, Vol. 33 Iss 4 pp. 265-269
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

Access to this document was granted through an Emerald subscription provided by 549136 []
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 www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
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.

 Feature Introduction

Cobots Collaborative robots ± ``cobots'' ± are in-

tended for direct interaction with a human
worker, handling a shared payload (Figure 1).
Michael Peshkin and
They are a marked departure from autono-
J. Edward Colgate mous industrial robots which must be isolated
from people for safety reasons. Cobots are
also distinct from teleoperators, in which a
human operator controls a robot and payload
remotely.
Cobots interact with people by producing
The authors software-defined ``virtual surfaces'' which
constrain and guide the motion of the shared
Michael Peshkin and J. Edward Colgate are both
payload, but add little or no power. Ergo-
Associate Professors of Mechanical Engineering at
nomic as well as productivity benefits result
Northwestern University, Evanston, Illinois, USA.
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

from combining the strength and computer-

interface of the cobot with the sensing and
Keywords
dexterity of the human worker.
Ergonomics, Robots This paper explains cobots as one approach
to an emerging class of materials handling
Abstract equipment called Intelligent Assist Devices
Collaborative robots ± ``cobots'' ± are intended for direct (IADs). We describe two cobots of this class
interaction with a human worker, handling a shared presently in industrial testbed settings. Future
payload. They are a marked departure from autonomous applications of cobots' virtual surfaces are
industrial robots which must be isolated from people for tool guidance in image guided surgery, and
safety reasons. Cobots are also distinct from teleopera- haptic display in which the surfaces of a CAD
tors, in which a human operator controls a robot and model can be felt.
payload remotely. Cobots interact with people by In 1995 Northwestern University and
producing software-defined ``virtual surfaces'' which General Motors Corporation initiated a pro-
constrain and guide the motion of the shared payload, but ject in the emerging area of Intelligent Assist
add little or no power. Ergonomic as well as productivity Devices (IADs) (Akella, 1999a). IADs are
benefits result from combining the strength and compu- intended to improve ergonomic working
ter-interface of the cobot with the sensing and dexterity conditions, product quality, and productivity
of the human worker. This paper explains cobots as one through an appropriate combination of ro-
approach to an emerging class of materials handling
botic technology with traditional Manual
equipment called Intelligent Assist Devices (IADs). We
Assist Devices (manually powered overhead
describe two cobots of this class presently in industrial
rail systems, jib cranes, etc.).
testbed settings. Future applications of cobots' virtual
Our particular approach to IADs was to
surfaces are tool guidance in image guided surgery, and
seek a way of implementing virtual surfaces,
haptic display in which the surfaces of a CAD model can
which we proposed as a primary form of
be felt.
interface between human worker and com-
puter in material handling tasks. Virtual
Electronic access
surfaces may be understood by analogy to the
The research register for this journal is available at role of a straightedge in drafting. Creating a
http://www2.mcb.co.uk/mcbrr/ir.asp straight line freehand is a task that most
The current issue and full text archive of this journal is people perform slowly and poorly. With the
available at help of a straightedge, however, the task is
http://www.emerald-library.com done better and quicker. Virtual surfaces are a
generalization of the straightedge: they pro-
vide physical guidance, in as many

Industrial Robot: An International Journal The authors wish to acknowledge the support of
Volume 26 . Number 5 . 1999 . pp. 335±341 the US National Science Foundation and the
# MCB University Press . ISSN 0143-991X General Motors Foundation.
335
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

Figure 1 A cobot and a human worker work together, with the cobot any of which can incapacitate a worker for
providing computer-generated ``virtual surfaces'' that the worker can make days to weeks. Cobots can address ergo-
use of in maneuvering a payload accurately into the door opening of a nomics concerns, and provide productivity
car on an assembly line benefits as well:
. Ergonomics: virtual surfaces produced by
cobots can guide worker and payload
motion, minimizing the need for the
worker to exert lateral or stabilizing forces
on a payload. Lateral and stabilizing
forces use the muscles of the upper body
and back, which are susceptible to injury.
Motive forces use principally the large
muscle groups of the lower body, with
less risk of injury. Thus we believe that
providing mechanical guidance may often
dimensions as necessary and along any shape be of greater ergonomic benefit than
path.
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

providing mechanical power.

Our invention of cobots early in the project . Navigation and inertia management: cobots
gave us a general and programmable means of can assist in the maneuvering of large,
setting up large-scale virtual surfaces. When a unwieldy objects, especially where com-
human worker pushes a payload up against a plicated motions or tight tolerances are
virtual surface, the payload's motion is con- necessary. In particular it is important to
fined to follow that surface, just as if it had realize that the virtual surfaces provided
run into a frictionless guide rail. When the by cobots are meant to be used inten-
payload is pulled away from the virtual tionally by the human worker. They are
surface, worker and payload motion is un- fixtures to be exploited, not barriers to be
constrained (``free mode''). avoided.
Conventional (servomotor-based) robots . Workspace isotropy (avoiding rail awk-
could in principle also be used to set up wardness): heavy assemblies supported by
virtual surfaces. Doing so, however, requires the very common passive overhead Car-
that the robot's joint motors be strong enough tesian rail systems are often much easier
to resist the worker's pushing forces and the to move along one translational axis than
payload's inertial forces that would penetrate the other. This non-uniformity of work-
a virtual surface. Such powerful motors pose space leads to awkward movements.
safety problems. Also, in practice, the per- Cobots mask the non-uniformity so that
formance of such virtual surfaces (e.g. their the payload's behavior seems intuitive to
smoothness) has proven to be poor. Cobots, the worker. We have built such a ``rail
by comparison, employ joints which me- cobot'', currently located at Ford Motor
chanically provide constraint. The constraint Company Advanced Manufacturing
is continuously adjusted under computer Technology Division (AMTD).
control, making it possible for the computer . Obstacle avoidance: virtual surfaces can be
to establish arbitrarily complex virtual sur- set up to surround and protect obstacles
faces. Virtual surfaces established by cobots in the workspace, effectively steering the
are intrinsically smooth, and do not require payload around them and preventing
powerful motors to defend them. collision.
. Software-driven material handling: cobots
can, via interface to a computer, guide an
Impact in materials handling operator in the correct selection of pickup
and placement locations as well as path.
Ergonomic problems in material handling As all aspects of the business world
have been widely recognized in recent years. undergo computerization, cobots can
Repetitive motions, excessive loads, awkward provide a computer interface to materials
postures and vibration can all lead to cumu- handling systems that must, for a variety
lative trauma disorders such as carpal tunnel of reasons, continue to use the special
syndrome, tennis elbow, and lower back pain, skills of human workers.
336
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

. Productivity: by exploiting virtual sur- that operator-generated forces could be

faces, operators can make quicker and distinguished from inertial forces of the pay-
surer movements. load.
The wheel is free to turn on its axle. There
How cobots use transmissions to create is no motor to drive its rolling motion. The
virtual surfaces ± example: the wheel is held vertical by a shaft whose axis is
``unicycle'' cobot coincident with the point of contact between
Cobots implement virtual surfaces by using wheel and rolling surface, i.e. there is no
transmissions. This avoids the need for brakes ``caster'' of the wheel. The wheel has a
or other dissipative elements, which necessa- steering angle bs defined as the angle of its
rily absorb energy of motion and thus do not rolling direction from the x axis. This angle is
provide a frictionless virtual surface. Trans- measured by a rotational encoder. Control of
missions are energetically neutral. A rolling the steering angular velocity os is accom-
wheel is the simplest example of a transmis- plished by a conventional velocity controller,
sion, and may be used to illustrate a simple which takes os as input. Owing to the absence
cobot. This is little more than a single wheel of caster, action of the steering motor cannot
in contact with a flat rolling surface, as shown
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

cause taskspace motion, and forces applied to

in the Figure. It has a two-dimensional
the handle do not create a torque on the
Cartesian (x-y) taskspace, parallel to the
steering motor. Thus there is a decoupling of
rolling surface. Using this model we will
taskspace motion from steering action.
discuss two essential behaviors of a cobot:
Also shown in the Figure is a rail system
virtual surface mode, and free mode.
which serves to keep the cobot upright and
The interface to the human operator is a
restrict it to its two-dimensional Cartesian
handle, shown just above a force sensor
taskspace. The rail system is instrumented
(Figure 2) which is able to measure the x-y
with translational encoders to measure the
forces applied by the operator. If there were a
position of the cobot within its taskspace.
payload, it would be below the force sensor so
Another rotational encoder monitors the
Figure 2 The simplest cobot is little more than a free- rolling speed of the wheel, u.
rolling wheel. By active control of the wheel's rolling This cobot, nicknamed Unicycle, is me-
direction, a control computer can confine the user's chanically well equipped to implement virtual
handle to follow a virtual path or surface. In ``free surfaces. In its two-dimensional taskspace, a
mode'' the computer can allow the operator to move virtual surface is a curve in the plane, p(s),
the device about freely, by steering the wheel to comply where s is path length along the curve and p is
with the operator's wishes transparently a 2-vector in the plane. Let us suppose we
wish this to be a bilateral surface to which the
cobot is to be confined, which we will call
path mode. In the next section we will address
its unrestricted motion when it is not in
contact with a virtual surface, free mode. A
unilateral virtual surface is accomplished by a
simple software switch between free mode
and path mode, based on whether the user's
applied force is directed toward the free side
or the prohibited side of the virtual surface.
In the absence of errors, confining the
cobot's motion to a curve in the plane
requires that we measure the cobot's position
s along the curve, and maintain its steering
angle bs such that its rolling direction is
tangent to the curve at that location. Thus
open-loop control for a cobot path-tracking a
virtual surface may be accomplished by os =
u/r where u is the measured rolling speed of
the wheel, r is the instantaneous radius of
curvature of the virtual surface, and os is the
337
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

commanded steering angular velocity. Closed Figure 3 Graduate student Witaya Wannasuphoprasit with a prototype
loop control is addressed in Colgate (1996). ``tricycle'' cobot capable of defending virtual surfaces in three-dimensions
The resulting virtual surface relies for its (x, y, and orientation y). A virtual surface (on screen) corresponds to the
strength and hardness not on actuators, but table edge
on the properties of a rolling wheel. The
wheel rolls freely despite large perpendicular
(``skidding'') loads, thus providing a low-
friction virtual surface. In practice we use
Rollerblade1 wheels, taking advantage of a
technology that has been optimized for a sport
requiring similar wheel properties. Virtual
surface strength well over 100 lbs is attain-
able. Perceptually, the virtual surface can be
easily confused with a well-greased rail,
confining motion to any programmed curve in
the plane.
Free mode is implemented by a servo loop
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

in which the operator's applied force is

measured by a force sensor, and the cobot's
single degree of freedom is steered to allow
motion in the direction that the operator's
force directs. To make the cobot feel like a
free mass, steering speed is inversely propor-
tional to rolling speed; thus when moving
quickly more user force is needed to turn.
Since any sort of software filter may be
applied to the operator's force, we are
afforded the opportunity for a great variety of
haptic effects.

An industrial materials handling cobot

Moving from the Unicycle cobot's two-
dimensional workspace (xy) to a three-di-
mensional planar workspace (xyy) requires a
Figure 4 General Motors' door-unloader cobot provides virtual surfaces
general theory of multidimensional cobot
which assure that the door does not collide with the car body, and which
control (Gillespie, 1999). We built an xyy
assists in several other aspects of the task. Yet the worker retains direct
``tricycle'' cobot, with the help of which
control of the payload as well
multidimensional cobot control was devel-
oped, while GM built an industrial version,
pictured in Figures 3 and 4.
With three wheels, virtual constraint sur-
faces can be defined in terms of orientation as
well as position within the plane. Kinemati-
cally, three wheels are one too many, and the
tricycle cobots can brake by intentionally
misaligning their wheels, such that their three
rolling axes do not coincide at a common
center of rotation. They can also display the
modes mentioned above: free mode in which
the wheels are steered so as to comply with
user forces; path mode in which the cobot
steers along a defined path through 3-space,
and virtual surface mode in which the wheels
are steered tangent to a software-defined
constraint surface, resisting user forces that
would violate the virtual surface.
338
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

A rolling wheel, which creates a nonholo- computer aided design (CAD). Haptic render-
nomic constraint, is essential to the cobots ing is analogous to ``graphical rendering'',
pictured. Each rolling wheel removes a degree referring to high-quality ``touchable'' represen-
of freedom from the basic mechanism. For tation of computer-generated surfaces.
instance, the tricycle's three degrees of planar As an example, consider the carved clay
freedom (x, y, y) are reduced to zero by its models used in conjunction with CAD
three wheels. In cobots, servo-control is used models in designing automobile bodies. Both
to selectively add apparent degrees of freedom time and information are lost in converting
by steering the wheels, so that (in free mode) between the two forms. Direct haptic ex-
the device appears to be unconstrained. In ploration of CAD models is desired, but the
contrast, conventional robots have multiple state of haptic rendering using conventional
mechanical degrees of freedom, which may be servo-actuated haptic displays is not yet
selectively reduced through servo-control to advanced enough to rival the sensitivity of
create apparent surfaces of constraints. In physical touch.
practice it is quite difficult for robots to Cobots may provide a superior technology
produce convincing virtual surfaces, but this for the haptic display of complex surfaces.
is the intrinsic mode of cobots. Cobots use a different physical principle to
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

Benefiting from our experience in designing display surfaces, relying on the mechanical
and programming the tricycle cobot, GM's properties of a continuously variable trans-
cobot is a rugged yet highly maneuverable mission (CVT) as the basis of virtual surfaces,
device. It assists in removing doors from rather than on powered servomotors. The
newly painted auto bodies prior to assembly CVT serves the same function in haptic-
of the cabin. This task was chosen because it display cobots as does the rolling wheel in the
was difficult for workers to remove the doors unicycle and tricycle cobots described above.
manually without marring the surfaces: the The CVT makes it possible to build cobots
curvature and styling of the body panels is having the familiar ``armlike'' design of many
such that a specific ``escape trajectory'' is robots. Figure 5 shows how CVTs make
needed to remove the door safely. Human
Figure 5 A cobot with the familiar workspace of a robotic ``arm''. The
versatility and dexterity are still very impor-
intended use is haptic display of virtual surfaces based on CAD solid
tant in other phases of the task; this is not a
models, so that a user can feel the surfaces of the object or parts designed.
task that ought to be fully automated.
Such cobots require revolute joints rather than rolling wheels as in the
The door-unloader cobot glides easily on
cobots shown in Figures 3 and 4. For this purpose they use the CVT shown
pairs of servo-steered RollerbladeTM wheels
in Figure 6
with only a few lbs of operator force. During
some task phases the operator controls posi-
tion while the cobot controls orientation,
aligning itself to the car body or to the door
rack across the walkway as appropriate. In
close-approach phase a virtual surface is
created close to the vehicle's rocker panel,
guiding the cobot to the correct location to
grip the door without colliding with the
vehicle. The ``escape trajectory'' is executed in
path mode. When crossing the walkway the
cobot is in free mode, giving the operator
direct and intuitive control over both trans-
lation and orientation.

Cobots for other applications

Haptic rendering
In the past decade, haptics has grown from an
unknown field to a widely recognized discipline
with both scientific and commercial signifi-
cance. Applications of haptics may be found in
entertainment, computer interface, image
guided surgery, training and rehabilitation, and
339
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

possible the revolute joints needed for ``arm- Figure 6 First prototype of a CVT which makes possible
like'' cobots. We are presently building a revolute-jointed (``armlike'') cobots. The internal
cobot (Figure 5) for haptic display, having a workings of more practical versions of the CVT are less
workspace similar to full human armspan. visible

Image guided surgery

In this application safety is essential, and a
cobot's ability to guide motion without
possessing a corresponding ability to move on
its own can totally remove concern about
some failure modes. Perhaps more impor-
tantly, the quality of a virtual surface enforced
by a cobot originates in its physical mechan-
ism, rather than in servo-controlled actuators,
thus yielding harder and smoother surfaces
than can be achieved by a robot. Preserving
the critical sense of touch in surgery requires
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

high quality ``shared control'' between sur-

geon and robot, for which smoothness of
motion is essential.
Rehabilitation and exercise
Popular weight training equipment, originally
designed for rehabilitation, uses shaped cams
and other mechanical components to confine
a user's limb or body motion to a particular
trajectory. While these trajectories are some-
what adjustable, far greater versatility could are a great number of CVT designs that hold
be achieved if the motion trajectories were two angular velocities in proportion, with o1/
encoded in software rather than frozen into o2 = c. We will express the transmission ratio
the mechanical design of the equipment. c as the tangent of an angle bs to maintain an
Cobots can enforce virtual trajectories with analogy to the rolling wheel. Just as for the
the smoothness, hardness, and safety required rolling wheel, for use in a cobot we require the
for this application. ability to rotate the transmission angle bs
without limit, through multiple revolutions.
Continuously variable transmission for We have developed a rotational CVT with
revolute-jointed cobots this behavior. Our first prototype is shown
The rolling wheel in the planar cobots may be above: the angular velocities of the rollers on
thought of as a translational transmission the left of the central sphere are held in a ratio
element. A transmission holds one velocity or determined by the angle of the rollers on the
angular velocity in proportion to another. The right. Understanding its kinematics from a
rolling wheel may be thought of as a device photo is nearly impossible. The kinematics of
which relates the x translational velocity of a this mechanism are described in detail in
certain point of a body to its y translational Peshkin (1996).
velocity, and holds those velocities in pro- CVTs have emerged as a crucial element of
portion. The proportion is adjustable by cobots. Strong, compact, and easily as-
steering the wheel: we have vy/vx = tan bs, sembled designs are needed. We have
where bs is the steering angle of the wheel. initiated a study of the contact mechanics of
Thus a steerable passive rolling wheel for- CVTs, which impacts torque-handling capa-
mally falls in the class of kinematic city, slip, and wear. We have also initiated a
mechanisms known as continuously variable program in CVT design.
transmissions, or CVTs (Figure 6).
For cobot architectures with revolute joints,
the wheel (as a CVT) is replaced by a device Conclusions
which holds two angular velocities in pro-
portion, with the ratio between them being Cobots hold great promise for human-com-
``steerable'' under computer control. There puter physical interaction. The philosophy
340
 Cobots Industrial Robot: An International Journal
Michael Peshkin and J. Edward Colgate Volume 26 . Number 5 . 1999 . 335±341

behind cobots is that shared control of Akella, P. (1999b), ``Cobots for the automobile assembly
motion, rather than amplification of human line'', in Prasad A., Peshkin, M., Colgate, E.,
power, is the appropriate metaphor for Wannasuphoprasit, W., Nagesh, N., Wells, J.,
Holland, S., Pearson, T. and Peacock, B.,
collaboration. Benefits may be expected in Proceedings of the IEEE 1999 International
ergonomics, in productivity, and in the inter- Conference on Robotics and Automation, Detroit,
face of computers and information systems to MI.
those many activities which continue to make Colgate, J.E. (1996), ``Nonholonomic haptic display'', IEEE
good use of uniquely human skills. ICRA Best Conference Paper award, in Colgate, J.E.,
Peshkin, M. and Wannasuphoprasit, W.,
Readers who wish to learn more about
Proceedings of the IEEE 1996 International
cobots may wish to consult the articles Conference on Robotics and Automation, Philadel-
mentioned in the references. Written material phia, PA.
of both greater and lesser technical detail may Gillespie, R.B. (1999), ``A general framework for cobot
be found at the cobot Web site, accessible via control'', in Gillespie, R.B., Colgate, J.E. and
the URL http://cobot.com. We have founded Peshkin, M., Proceedings of the 1999 IEEE
International Conference on Robotics and
a company for the commercialization of cobot
Automation, Detroit, MI.
technology, Collaborative Motion Control Moore, C.A. (1999), ``Design of a 3R cobot using
(``CoMoCo'') Inc., which is also reachable at continuously variable transmissions'', in Moore,
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)

the above URL. Some intellectual property C.A., Peshkin, M.A. and Colgate, J.E., Proceedings
rights in cobots are held by Northwestern of the IEEE 1999 International Conference on
University. Robotics and Automation, Detroit, MI.
Peshkin, M. (1996), ``Passive robots and haptic displays
based on nonholonomic elements'', in Peshkin, M.,
Colgate, J.E. and Moore, C., Proceedings of the IEEE
References 1996 International Conference on Robotics and
Automation, Philadelphia, PA.
Akella, P. (1999a), ``Intelligent Assist Devices: a new Wannasuphoprasit, W. (1998), ``Cobots: a novel material
generation of ergonomic tools (workshop)'', 1999 handling technology'', in Wannasuphoprasit, W.,
International Conference on Robotics and Akella, P., Peshkin, M. and Colgate, J.E.,
Automation, Detroit, MI. Proceedings of the 1998 ASME IMECE.

341
 This article has been cited by:

1. Mustafa Suphi Erden, Bobby Marić. 2011. Assisting manual welding with robot. Robotics and Computer-Integrated
Manufacturing 27:4, 818-828. [CrossRef]
2. Suwoong Lee, Susumu Hara, Yoji Yamada. 2010. A Safety Measure for Control Mode Switching of Skill-Assist for Effective
Automotive Manufacturing. IEEE Transactions on Automation Science and Engineering 7:4, 817-825. [CrossRef]
3. Pablo Gonzalez de Santos, E. Garcia, Javier Sarria, Roberto Ponticelli, Jesus Reviejo. 2010. A new manipulator structure for
power‐assist devices. Industrial Robot: An International Journal 37:5, 452-458. [Abstract] [Full Text] [PDF]
4. M.S. Erden, T. Tomiyama. 2010. Human-Intent Detection and Physically Interactive Control of a Robot Without Force
Sensors. IEEE Transactions on Robotics 26:2, 370-382. [CrossRef]
5. Bibliography 549-602. [CrossRef]
6. Aslam Pervez, Jeha Ryu. 2008. Safe physical human robot interaction-past, present and future. Journal of Mechanical Science
and Technology 22:3, 469-483. [CrossRef]
7. Dongjun Lee, P.Y. Li. 2003. Passive bilateral feedforward control of linear dynamically similar teleoperated manipulators.
IEEE Transactions on Robotics and Automation 19:3, 443-456. [CrossRef]
Downloaded by ONDOKUZ MAYIS UNIVERSITY At 10:58 05 November 2014 (PT)