Abstract—The design of master manipulators for master-slave

surgical robotic systems is important because it may influence

slave manipulator performance as well as the operator’s

workload. However, no design strategy has been presented thus

far for optimizing the master manipulator design parameters. A

master manipulator prototype and an experimental setup were

developed for investigating design parameter influence using our

master-slave microsurgical robotic system. The preliminary

results showed that the relative position of the holding point, the

corresponding point for the slave manipulator’s working point,

and the center of the gimbals are important for master

manipulator design, especially for tasks requiring large or

frequent posture changes. The experimental results also

suggested that the optimized parameters would contribute to

enhancing task efficiency and decreasing the workload, rather

than increasing task accuracy.

I. INTRODUCTION

ASTER –slave robotic systems have been employed in

surgical robotics for enhancing surgeons‟ dexterity as

well as enabling motion scaling. The most famous surgical

robotic system, the “da Vinci Surgical System,” has its own

master manipulator design, facilitating intuitive manipulation

of the slave robotic arms. Many research groups have also

tried to design their original master manipulators or used

available haptic devices [1]; however, to the best of the

present authors‟ knowledge, there is no design strategy for

surgical master manipulators. Master manipulator design

affects system usability, the slave manipulators‟ motion

accuracy, and operator workload. Further, it is obvious that

better master manipulator design facilitates the introduction of

surgical robots to clinical cases. Moreover, master

manipulator design may depend on the target clinical

applications or tasks, and thus, a design theory is required.

The research focus in laparoscopy has been on the

enhancement of dexterity and usability as well as the

implementation of haptics, rather than the enhancement of

accuracy, and thus, many research groups use commercial

haptic devices such as the PHANTOM series (SensAble,

Manuscript received on January 31, 2012.

Y. Kamei, S. Tanaka, K. Harada, Y. Baek, Y. Ida, N. Sugita and M.

Mitsuishi are with the Department of Mechanical Engineering, The

University of Tokyo, Bunkyo-ku, Tokyo 113-8656 Japan (Tel. and Fax:

+81-3-5841-6357; e-mail: kanako@ nml.t.u-tokyo.ac.jp).

S. Sora is with the Department of Neurosurgery, Tokyo Metropolitan Police

Hospital, Nakano-ku, Tokyo 164-8541 Japan.

A. Morita is with the Department of Neurosurgery, NTT Medical Center

Tokyo, Shinagawa-ku, Tokyo 141-8625 Japan.

USA) (for example, in [2], as shown in Fig. 1(a)) and the

Delta/Omega/Sigma series (Force Dimension, Switzerland)

(for example, in [3], as shown in Fig. 1(b)). There are also

several original designs using the Delta structure [1, 4].

Meanwhile, accuracy and precision are of the utmost

importance in microsurgery. For example, microsurgical

robots for neurosurgery need to perform very fine and

complex tasks such as the anastomosis of 0.7 mm blood

vessels. Although there are several neurosurgical robotic

systems [5-8], their target applications are limited to simple

tasks such as needle insertion and tumor removal. Hence, the

usability of master manipulators for complicated tasks

remains to be established. Microsurgical robots include eye

surgery robots (for example, the eye surgery robot reported in

[9]). However, no complex tasks require to be performed in an

eye surgery. Consequently, we have focused on the usability

of a master manipulator, especially for micro neurosurgical

robots having six degrees of freedom (DoFs), with the three

positioning DoFs decoupled from the three orientational

DoFs.

In this work, the design parameters of surgical master

manipulators were investigated using a microsurgical robotic

system in order to clarify the unique microsurgery features

that need to be considered in the design of master-slave

surgical robots. Master manipulator prototypes were designed

by attaching a mechanism to a commercial haptic device in

order to vary the design parameters. Experiments were

conducted in order to quantify the effects of design parameters

on the robot performance and workload.

The paper is organized as follows. The prototype design is

described in Section II, and Section III details the

experimental setup and method. Section IV reports the

experimental results, and the paper is concluded in Section V.

(a) (b)

(c) (d)

Fig. 1. Master manipulators for surgical robotic systems. (a) Raven [2], (b)

DLR MiroSurge [3], (c) NeuroArm [5], and (d) Brain Tumor Removal Robot

System [6].

Study on Master Manipulator Design Parameters

for Robotic Microsurgery

Y. Kamei, Student Member, IEEE, S. Tanaka, Student Member, IEEE, K. Harada, Member, IEEE,

Y. Baek, Student Member, IEEE, Y. Ida, S. Sora, A. Morita, N. Sugita, and M. Mitsuishi, Member, IEEE

M

The Fourth IEEE RAS/EMBS International Conferenceon Biomedical Robotics and BiomechatronicsRoma, Italy. June 24-27, 2012

978-1-4577-1198-5/12/$26.00 ©2012 IEEE 847

II. PROTOTYPING

A. Master-Slave Robotic System for Microsurgery

Master manipulator designs were investigated using our

master-slave microsurgical robotic system [10]. This system

was designed for both neurosurgery and eye surgery. The

anastomosis of 0.3 mm artificial blood vessels [11], which is a

very difficult manual operation, was demonstrated with the

system.

Figure 2 shows the master-slave microsurgical robotic

system overview. The operator (i.e., the surgeon) manipulates

two master arms while looking at the microscopic stereo view

provided on the 3D monitor. The slave robotic system is

composed of two robotic arms, each with six DoFs, and a

surgical tool unit with one gripping DoF is mounted onto each

arm. The surgical unit can be interchanged with another unit

based on the target clinical application or task. The

translational motion of the master manipulator is scaled down

and transmitted to the slave robot.

Microscope

Forceps

Orientation Component

Translation Component

SurgicalField

Microscopic View

3D monitor

Master manipulators

Stereo microscope

Surgical tool unit

Slave

manipulators

Surgical field

Microscopic view

Fig. 2. Master-slave microsurgical robotic system.

Human arm motion was previously quantified and the

master manipulators shown in Fig. 2 were designed [12]. The

master manipulator employed a pen-like gripper because the

manner in which a pen is held, called a tripod grip, is similar to

the manner in which a pair of surgical tweezers is held in

microsurgery. Figure 3 (a) shows the manner in which a pair

of tweezers is held, where C is the working point, H is the

holding point, and O is the approximate center of rotation. The

gripper of the master manipulator (Fig. 3(b)) was designed to

replicate these relative positions; the hand grasping point, H,

was approximately 30 mm from the gimbal center, O, whereas

the point, C, to which the slave robot‟s tip corresponded was

set 80 mm from the gimbal center, O.

H

C

O

H C

O

(a) (b)

Fig. 3. The manner in which a pair of surgical tweezers is held. C is the

working point, H is the holding point, and O is the center of revolution.

Surgeons were not satisfied with the usability of the master

manipulators, regardless of the similarity of the gripping pose,

although the dexterity was sufficient to perform a complicated

task with high precision. It was also difficult for us to identify

which design parameters to change because the design process

was not systematic. Therefore, the design process was

reviewed. In many master manipulators reported in literature,

the working point of a surgical tool, xc′, (Fig. 4(a)) is

correlated to the center of the gimbals, O, (Fig. 4(b)) of the

master manipulator. For example, in the x-coordinate whose

origin was set at the gimbal center (Fig.4 (b)), the

corresponding slave manipulator position coincides with the

origin (xc = 0) and the surgeon holds the proximal side of the

gripper (xh < 0). Both the holding and corresponding points

were set at the distal side of the gripper (xc > xh > 0) in our

previous master manipulator design, as shown in Fig.4 (c). It

seems that the configuration shown in Fig. 4 (b) is more

intuitive, whereas the configuration in Fig. 4(c) was designed

to replicate the actual holding position used in microsurgery.

Therefore, the ways in which the similarity to the actual

surgical situation may affect the performance and workload in

a robotic microsurgery scenario were investigated. A

prototype of the master manipulator was designed in order to

quantify the influence of the design parameters xc and xh.

xc‟

x

xc

O

xh

xxc

xh

O

(a) (c)(b)

Fig. 4. Master-slave position correspondence. (a) Working point of the

surgical tool mounted on the slave manipulator, (b) Corresponding and

holding points of the master manipulator in [5], and (c) Corresponding and

holding points of the master manipulator of the master-slave microsurgical

robotic system [12].

B. Master manipulator prototype design

A PHANTOM Omni device (SensAble, USA) was used to

measure the positioning DoFs (three DoFs) and a gimbal

mechanism with a stylus was attached in order to measure the

orientation DoFs (three DoFs), as shown in Fig. 5. The

original gimbals of the PHANTOM Omni were glued and

fixed. The advantage of using a commercial device is that the

results can be shared with the research community and used

for the design of a new manipulator for a master-slave surgical

robotic system.

Three miniature encoders with a resolution of 0.0225°

(MES-9-1000PST16E, Microtech Laboratory Inc., Japan)

were implemented in the gimbals and provided higher pose

measurement accuracy in comparison to the original

PHANTOM Omni gimbals. This was necessary in order to

avoid the effect of the relatively low angular resolution of the

original gimbals on slave manipulator performance. The

pen-like stylus was attached to the gimbal structure, which

848

provides more configurable holding point positions. The

holding point, xh, is a mechanical design parameter, whereas

the corresponding point, xc, is a parameter to be set in the

control program.

Fig. 5. Master manipulator for the experiment. (a) PHANTOM Omni

(SensAble, USA), (b) Gimbals with a stylus.

III. EXPERIMENTS

A. Experimental Setup

The experimental setup is shown in Fig. 6 and the system

overview is shown in Fig. 7. The experiments were

performed for the right hand only because all subjects were

right-handed and used their right hands for needle

manipulation in microsurgery. The Libralis library [13] was

used for PHANTOM gravity compensation. The

communication between the master and slave systems was

updated at 100 Hz and the servo rate of the slave robotic

system was 1 kHz. The response was sufficient, and it was

assumed that slave performance and operator workload

were not influenced by the mechanical setup or control.

3D monitor

Stylus

Gimbals

PHANTOM

Omni®

Arm rest

Stereo Microscope

Slave

manipulator

(a) (b)

Fig. 6. Experimental setup. (a) Master system (b) Slave system.

3D monitor

RTLinux UDPWindows

100 Hz

Slave

manipulator

Stereo microscope Master

manipulator

Foot switch

FireWire USB

Pulse counter

Slave

system

Master

system

Arm rest

Fig. 7. Experimental setup: system configuration.

B. Design of Experiments

The microscopic magnification was set such that a 10 mm

circle could fit in the view. A 27 G needle with its tip painted

red was attached to the distal end of the surgical tool unit, and

the position of the needle was estimated and tracked using the

image processing method described in the next subsection.

The position of the tip of the needle was set to coincide with

that of the remote center of motion of the slave manipulator.

Three tasks were designed, namely, (1) a Tracing Task, (2)

a Pointing Task, and (3) a Posture-changing Task, as

illustrated in Fig. 9. The tasks were developed to contain some

of the motion elements required in microsurgical tasks. The

motion scaling ratio was set at 3x, which reflected a

neurosurgeon‟s opinion that the slave manipulators can be

intuitively maneuvered at the ratio in neurosurgical tasks. The

distance between the gimbal center and holding point Xh = –xh

and the distance between the corresponding point and holding

point Xc = xc xh were varied in the experiment, as shown in

Fig. 8. Twelve pairs of parameters (Xh = 20, 40, 60 mm and Xc

= 30, 0, 20, 60 mm) were tested in the preliminary

experiment. Two sets of twelve pairs were tested in random

order for each subject. Each subject was asked to keep his

forearm on the arm rest. In each task, the subject was asked to

provide a Subjective Score ranging from 1 (uncomfortable) to

10 (comfortable).

1) Tracing Task

This task was designed to examine the usability of the

master manipulator in the general position-changing

maneuvers. The subject traced a circle with a diameter of

10 mm in the tracing task. The circle was printed on a

piece of paper and the subject was asked to place the

needle tip as close as to the paper as possible. The task

completion time, RMS error, and the Subjective Score

were evaluated.

2) Pointing Task

The purpose of this task was to examine the usability in

precise targeting. The subject placed the needle tip in

small circles of 0.3 mm in diameter in the order of 0, 1, 0,

2, 0, 3, 0, 4, 0, as shown in Fig. 9 (b). The task

completion time, length of the trajectory, task

completion ratio, and Subjective Score were evaluated.

The trajectory length in 3D space was calculated from

the position of the corresponding master manipulator

point. Task completion evaluation success was defined

as the precise placement of the needle tip within each

small circle.

3) Posture-changing Task

This task was selected to evaluate the usability in

changing the slave manipulator posture while

maintaining the tip position. The subject aligned the

needle with the horizontal line starting from the line with

an angle of 45°, as shown in Fig. 9, and placed it back in

the original posture while keeping the needle tip at the

center of the cross. This set was repeated three times. The

task completion time, RMS error, and the Subjective

Score were evaluated.

(a) (b)

849

O

C

H

x Xc

Xh

Fig. 8. Coordinate system setting for the experiment.

10 mm

(a) (b) (c)

4

1

3

0

2

Fig. 9. Experimental tasks. (a) Tracing Task, (b) Pointing Task, and (c)

Posture-changing Task.

C. Tool tip detection

Image-based tool tip detection was developed and used to

obtain the position of the tip of the needle. The tracked

position was used to calculate the RMS error in the three

tasks mentioned in the previous subsection. As mentioned

above, the needle tip was painted red to simplify the

required image processing. It was easy to extract the red

color region in the HSV image space (consisting of hue,

saturation, and intensity) in the designed experimental setup

because the background color was white. The extracted red

colored region was converted to a binary image and the tool

tip was estimated by detecting the upper and lower contours

of the region.

Figure 10 shows an example of tool tip detection. The

binary image was generated by extracting the red colored

region in the microscopic image using predefined thresholds

for the hue and saturation values. The intensity value was

not used in the experiment because it was prone to error

because of the specular effect observed in the obtained

microscopic image. Next, several sets of two points were

aligned to the upper and lower contours of the extracted

region with a preset interval in order to find the lateral

contours of the needle. Thereafter, upper and lower contour

lines were generated by calculating the coefficients based on

the least square estimation. The tool tip point was defined as

the left-most point of the region located on the centerline of

the two contour lines. The tip position was tracked at a rate

of 30 Hz, with measurement accuracies of <50 m RMS.

Although there remains scope for improvement in the

detection accuracy, it was assumed to be sufficient for

master manipulator design parameter analysis.

Fig. 10. Example of tool tip detection: (a) input microscopic image, (b)

binary image of the red-colored tool region, (c) control points for the upper

and lower contours, and (d) line detection and tool tip estimation.

D. Statistical Analysis

The statistical significance of the differences in the

parameter sets was analyzed by a repeated-measures analysis

of variance (ANOVA) with two within-subject factors (Xh and

Xc), followed by post hoc analysis (Sidak‟s multiple

comparison test based on the estimated marginal means). The

statistical difference was accepted at p < 0.05. The

Greenhouse-Geisser method was used to adjust the degrees of

freedom, where appropriate. All statistical analyses were

conducted using Version 20 of SPSS Statistics (IBM, USA).

IV. RESULTS

A. Subjects

Two medical doctors and eight Engineering students

participated in the experiments.

B. Tracing Task

The ANOVA revealed significant main effects of factor Xc

on the Subjective Score (F(3, 27) = 7.313, p < 0.05), whereas

the main effects of factor Xh and the interaction Xh Xc on the

Subjective Score were not significant. The Sidak-corrected

pairwise comparisons revealed higher Subjective Scores for

Xc = 0 or 20 mm than for Xc = 30 mm (p < 0.05) (Fig. 11, left).

With regard to the RMS error, only the main effects of factor

Xh were significant (F(2, 18) = 4.456, p < 0.05). However, no

significant differences were revealed after multiple

comparisons of the error in each value of Xh, (Fig.11, right).

With respect to the task completion time, none of the main

effects of the factors was detected (data not shown).

1

2

3

4

5

6

7

8

9

10

-30 0 20 60

Su

bje

cti

ve S

co

re

Xc (mm)

*

*

0

50

100

150

200

250

300

350

400

450

500

20 40 60

RM

S E

rro

r (

m)

Xh (mm) Fig. 11. Tracing task result. Left: Estimated marginal means of the

Subjective Score for Xc. Bars indicate the standard errors. The asterisk (*)

indicates a pair with a significant difference (p < 0.05). Right: Estimated

marginal means of the RMS Error for Xh.

C. Pointing Task

The ANOVA revealed significant main effects of factors Xc

(F(1.476, 13.281) = 10.222, p < 0.05) and Xh (F(2, 18) =

4.868, p < 0.05) on the Subjective Score, whereas the main

effects of the interaction Xh Xc were not significant. Pairwise

comparisons revealed higher Subjective Scores for Xc = 0 and

60 mm than for Xc = 30 mm (p < 0.05) (Fig. 12, left). It was

also revealed that the Subjective Score for Xh = 20 mm was

higher than that for Xh = 60 mm (p < 0.05) (Fig. 12, right).

With regard to the task completion time, the significant main

effects of the factors Xc (F(3, 27) = 15.656, p < 0.001) and Xh

(a) (b) (c) (d)

850

(F(2, 18) = 6.755, p < 0.05) were revealed, whereas the main

effects of the interaction Xh Xc were not significant. Multiple

comparisons revealed shorter task completion times for Xc = 0,

20, and 60 mm in comparison to those for Xc = 30 mm (p <

0.05) (Fig. 13, left). Additionally, a shorter task completion

time was observed for Xh = 40 mm compared with Xh = 60 mm

(p < 0.05) (Fig. 13, right). Concerning the task completion rate

(Fig. 14), the significant main effects of factor Xc (F(3, 27) =

3.035, p < 0.05) were revealed, whereas no other main effects

were detected. The subsequent comparison test did not detect

any statistically significant differences in task completion

rates for Xc. No significant main factor effects were detected in

the lengths of the trajectories.

1

2

3

4

5

6

7

8

9

10

-30 0 20 60

Su

bje

cti

ve

Sco

re

Xc (mm)

**

1

2

3

4

5

6

7

8

9

10

20 40 60

Su

bje

cti

ve

Sco

re

Xh (mm)

*

Fig. 12. Subjective Score for the pointing task. Left: Estimated marginal

means of the Subjective Score for Xc. Right: Estimated marginal means of the

Subjective Score for Xh. Bars indicate the standard errors. The asterisk (*)

indicates a pair with a significant difference (p < 0.05).

0

5

10

15

20

25

30

-30 0 20 60

Ta

sk

Co

mp

leti

on

Tim

e (

s)

Xc (mm)

**

*

0

5

10

15

20

25

30

20 40 60

Ta

sk

Co

mp

leti

on

Tim

e (

s)

Xh (mm)

*

Fig. 13. Task completion time for the pointing task. Left: Estimated marginal

means of the task completion time for Xc. Right: Estimated marginal means

of the task completion time for Xh. Bars indicate the standard errors. The

asterisk (*) indicates a pair with a significant difference (p < 0.05).

0

10

20

30

40

50

60

70

80

90

100

-30 0 20 60

Task C

om

ple

tio

n R

ate

(%

)

Xc (mm) Fig. 14. Task completion rate of the pointing task. Estimated marginal

means for task completion rate according to the value of Xc. Bars indicate the

standard errors. No significant differences are observed.

D. Posture-changing Task

The ANOVA revealed significant main effects of factor Xc

(F(3, 27) = 7.293, p < 0.05) and the interaction Xh Xc (F(6,

54) = 7.451, p < 0.05) on the Subjective Score, whereas the

main effects of factor Xh were not significant. Fig. 15 (left)

shows the estimated marginal means for Xc at three Xh levels,

whereas Fig. 15 (right) shows the estimated marginal means

for Xh at four Xc levels. Regarding the task completion time,

the main effects of factor Xc (F(1.439, 12.951) = 6.696, p <

0.05) and the interaction Xh Xc (F(2.584, 23.255) = 3.831, p

< 0.05) were significant, whereas no significant main effects

were detected for factor Xh. The results of the multiple

comparison tests are shown in Fig. 16.

1

2

3

4

5

6

7

8

9

10

-30 0 20 60

Su

bje

cti

ve S

co

re

Xc (mm)

20 40 60

*

Xh (mm)

*

*

1

2

3

4

5

6

7

8

9

10

20 40 60

Su

bje

cti

ve S

co

re

Xh (mm)

-30 0 20 60** *

**

Xc (mm)

Fig. 15. Subjective Score for the posture-changing task. Left: Estimated

marginal means of the Subjective Score for Xc. Right: Estimated marginal

means of the Subjective Score for Xh. Bars indicate the standard errors. The

asterisk (*) indicates a pair with a significant difference (p < 0.05).

0

5

10

15

20

25

30

-30 0 20 60

Task C

om

ple

tio

n T

ime (

s)

Xc (mm)

20 40 60

Xh (mm) *

0

5

10

15

20

25

30

20 40 60

Task C

om

ple

tio

n T

ime (

s)

Xh (mm)

-30 0 20 60

Xc (mm)

** *

Fig. 16. Task completion time for the posture-changing task. Left: Estimated

marginal means of the task completion time for Xc. Right: Estimated

marginal means of the task completion time for Xh. Bars indicate the standard

errors. The asterisk (*) indicates a pair with a significant difference (p <

0.05).

E. Summary of Experimental Results

Table I summarizes the experimental results, including the

results that are not explained in detail in the previous

subsection.

The performance of the slave manipulator in terms of

accuracy was not influenced by the relative position of the

holding point, corresponding point, and the center of the

gimbals. The efficiency of the tasks, which was measured as

the task completion time or trajectory length, was influenced

by the relative positions for the pointing task and the posture

changing task, and this was probably because both tasks

required bigger or more frequent posture changes, which can

851

be more affected by the master manipulator design. The

Subjective Score was influenced in most cases.

The result of the preliminary experiment suggested that the

accuracy of the slave robot could be somehow independent of

the mechanical design of the master manipulator, though the

efficiency of the task and the operator workload could be

influenced by the design parameters, especially for tasks

requiring large or frequent posture changes.

Although more data should be collected for further

investigation of the optimal design parameters, it was clarified

that the relative position of the holding point, corresponding

point, and the center of the gimbals are important for master

manipulator design for microsurgery, especially for enhancing

usability.

TABLE I Summary of the experiments: The significant influence of the

parameters on each task; (a) the corresponding point should coincide with the

holding point or be located at its distal side; (b) the holding point should be

closer to the gimbal center; (c) the influence is proved, although the suitable

parameter setting remains unclear; (d) the corresponding point should

coincide with the holding point or be placed closer to the holding point at its

distal side.

1) Tracing 2) Pointing 3) Posture-changing

N Y (b) Y (c)

Efficiency N Y (b) Y (c)

Accuracy N N N

Y (a) Y (a) Y (d)

Efficiency N Y (a) Y (d)

Accuracy N N N

X h

Subjective Score

Performance

X c

Subjective Score

Performance

V. CONCLUSION

We developed a master manipulator prototype in order to

investigate the master manipulator design parameters for

microsurgery. The preliminary results showed that the relative

position of the holding point, corresponding point, and the

center of the gimbals are important for master manipulator

design for microsurgery, especially for enhancing usability.

The future work will include experiments with greater

parameter variance in order to systematically define the

design parameters.

ACKNOWLEDGMENT

This study was partially supported by the Global Center of

Excellence for Mechanical Systems Innovation (GMSI) at the

University of Tokyo and a Grant-in-Aid for Scientific

Research (S) 23226006.

REFERENCES

[1] J. Arata, H. Kondo, M. Sakaguchi, and H. Fujimoto, “Development of a

haptic device „DELTA-4‟ using parallel link mechanism,” in

International Conference Robotics and Automation (ICRA), pp.

294-300, 2009.

[2] G. Sankaranarayanan et al., “Portable surgery master station for mobile

robotic telesurgery,” in Proceedings of the 1st International

Conference on Robot Communication and Coordination, p. 28, 2007.

[3] U. Hagn et al., “DLR MiroSurge: a versatile system for research in

endoscopic telesurgery.,” International journal of computer assisted

radiology and surgery, vol. 5, no. 2, pp. 183-93, Mar. 2010.

[4] K. Tadano and K. Kawashima, “Development of a master slave system

with force sensing using pneumatic servo system for laparoscopic

surgery,” in Robotics and Automation, pp. 10-14, 2007.

[5] A. D. Greer, P. M. Newhook, and G. R. Sutherland, “Human–machine

interface for robotic surgery and stereotaxy,” Mechatronics,

IEEE/ASME Transactions on Mechatronics, vol. 13, no. 3, pp.

355-361, 2008.

[6] M. Takagi, T. Wada, J. Arata, Y. Hayashi, Y. Kajita, and H. Fujimoto,

“Development of Surgical Cockpit for Brain Tumor Removal Robot

System,” in the JSME Conference on Robotics and Mechatronics,

2011.

[7] K. Hongo, Y. Kakizawa, and J. Koyama, “Microscopic-manipulator

system for minimally invasive neurosurgery: preliminary study for

clinical application,” International Congress Series, vol. 1230, pp.

275-280, 2001.

[8] P. Eldridge, “Use of the NeuroMate stereotactic robot in a frameless

mode for functional neurosurgery,” The International Journal of

Medical Robotics and Computer Assisted Surgery, pp. 107-113, 2006.

[9] H. C. M. Meenink, R. H. P. C. J. N. Rosielle, M. Steinbuch, H.

Nijmeijer, and M. C. D. Smet, “A master-µslave robot for vitreo-retinal

eye surgery,” in The Euspen International Conference, Delft, 2010.

[10] Y. Ida, N. Sugita, T. Ueta, Y. Tamaki, K. Tanimoto, and M. Mitsuishi,

“Microsurgical robotic system for vitreoretinal surgery,” International

journal of computer assisted radiology and surgery, vol. 7, no. 1, pp.

27-34, Jan. 2012.

[11] Y. M. Baek et al., “Highly precise master-slave robot system for super

micro surgery,” 2010 3rd IEEE RAS & EMBS International

Conference on Biomedical Robotics and Biomechatronics, pp.

740-745, Sep. 2010.

[12] H. Takahashi, T. Yonemura, N. Sugita, and M. Mitsuishi, “Master

manipulator with higher operability designed for micro neuro surgical

system,” IEEE International Conference on Robotics and Automation

(ICRA 2008), pp. 3902-3907, 2008.

[13] A. Formaglio, M. Fei, S. Mulatto, M. de Pascale, and D. Prattichizzo,

“Autocalibrated gravity compensation for 3DoF impedance haptic

devices,” Haptics: Perception, Devices and Scenarios, pp. 43-52,

2008.

852