Facultatea Constructii de Masini

Specializarea: Robotica Engleza

Master anul II

NEMESY
Robot for neurosurgery

Student:
Molitorisz Andor
 Table of Contents

1. Introduction to robots in medical applications ......................................................................................... 3

1.1 General information about medical robots ......................................................................................... 3
1.2 The fundamental requirements from a medical robot ........................................................................ 4
1.3 Research Directions ............................................................................................................................ 5
2. Robotic Neurosurgery .............................................................................................................................. 7
2.1 Developments of robotic neurosurgery .............................................................................................. 7
2.2 The Next Step in Robotic Neurosurgery .......................................................................................... 15
3. NEUMESY Neurosurorgical Robot ....................................................................................................... 18
3.1. The initial experiment description ................................................................................................... 18
3.2 Medical and Robotics requirements ................................................................................................. 21
3.3 Robot kinematics .............................................................................................................................. 23
3.4 Conclusions ........................................................................................................................................ 2
References .................................................................................................................................................... 3
 1. Introduction to robots in medical applications

1.1 General information about medical robots

The first generation of surgical robots is already being installed in a number of operating
rooms around the world. Robotics is being introduced to medicine because it allows for
unprecedented control and precision of surgical instruments in minimally invasive procedures.
So far, robots have been used to position an endoscope, perform gallbladder surgery and correct
gastroesophogeal reflux and heartburn. The ultimate goal of the robotic surgery field is to design
a robot that can be used to perform closed-chest, beating-heart surgery. The use of robotics in
surgery will expand over the next decades without any doubt.

In 1979, the Robot Institute of America, an industrial trade group, defined a robot as “a
reprogrammable, multifunctional manipulator designed to move materials, parts, tools, or other
specialized devices through various programmed motions for the performance of a variety of
tasks.” Such a definition leaves out tools with a single task (e.g., stapler), anything that cannot
move (e.g., image analysis algorithms), and nonprogrammable mechanisms (e.g., purely manual
laparoscopic tools). As a result, robots are generally indicated for tasks requiring programmable
motions, particularly where those motions should be quick, strong, precise, accurate, untiring,
and/or via complex articulations. The downsides generally include high expense, space needs,
and extensive user training requirements. The greatest impact of medical robots has been in
surgeries, both radiosurgery and tissue manipulation in the operating room, which are improved
by precise and accurate motions of the necessary tools. Through robot assistance, surgical
outcomes can be improved, patient trauma can be reduced, and hospital stays can be shortened,
though the effects of robot assistance on long-term results are still under investigation. Medical
robots have been reviewed in various papers since the 1990s. Many such reviews are domain-
specific, for example, focusing on surgical robots, urological robots, spine robots, and so forth.

Robotic–assisted surgery is a new trend in medicine, which aims to help the surgeon by
taking advantage of robots’ high accuracy and accessibility. Introducing a robotic assistant as an
integral part of the surgical tool array provides the surgeon with several advantages.

These advantages include off-loading of the routine tasks and reduction of the number of
human assistants in the operating room. In addition, by using the capabilities of the robot, the
surgeon can complement his own skills with the accuracy, motion steadiness, and repeatability
of the robot. The experimental comparison, presented in compared the performance of a human
assistant and a robotic assistant in manipulating a laparoscope.

The approaches to robot assisted surgery divide into three main approaches. These
approaches are the active execution approach, the semi-active approach, and the passive
approach. In the active execution approach, the robot actively performs surgical procedures such
as bone cutting and milling.

Among the vast array of robotic structures, the parallel structure seems promising
because of its advantages that fit medical applications. Therefore, some investigators focused on
exploring the capabilities of parallel robots in medical applications

1.2 The fundamental requirements from a medical robot

In order to insure the success of a medical robot, four fundamental requirements must be
fulfilled. The first and most crucial requirement is safety. The following seven criteria constitute
the safety requirement:

1) Effective control: The robot must allow, in all configurations, effective control ofthe tool with
both speed and force control schemes implemented.

2) Limited Workspace: The robot must have limited workspace in order to preventhazardous
collisions between its moving parts and the medical staff or the patient.

3) Limited Forces or Force feedback: In applications where the robot is active in performing
surgical procedures that include tactile tasks, the force applied by the tool must be limited.
Alternatively, in applications where the robot acts as a slave, the robot must convey a maximum
amount of data to the surgeon about the forces exerted on the tool. This requirement is essential
in the process of bone cutting where different levels of force are required during different stages
of the cut.

4) Immunity against magnetic interference of other surgical tools.

5) Full control option: In applications where the robot performs automated tasks, the control
program must allow the surgeon, in any stage in the task, to interrupt the automatic execution
process and take over the control to his hands.
6) Fail safe features: The most reliable systems will inevitably fail in some stage of their service.
Based on this premise the robot must support a fail-safe mode. This includes keeping the
position of the tool when the power supply is lost, electrical limiting of the end effector’s speed
and force even when the control program fails.

7) Safe behavior near singular configurations: The path planning of the robot should avoid
passing near singular configurations. However, in the cases where the robot acts as a slave, the
surgeon might manipulate it into singular configurations. Therefore, the architecture of the robot
must provide signals for the surgeon that warn him from approaching singularity.

The second requirement from a medical robot is compactness in size and lightness. This
ensures that the robot does not consume a large amount of essential space in the operating room
and facilitates the relocation of the robot in different positions for different tasks. The third
requirement is simple operation in order to improve the learning curve of new surgeons. The last,
but not least important, requirement is the requirement for easy sterilization. This requirement is
critical since any tool in the operating room must either be sterilized or covered with sterile
drapes in order to prevent infections. [1]

1.3 Research Directions

Medical robotics is a young and relatively unexplored field made possible by technical
improvements over the past couple of decades. Currently available systems have been available for too
short time to allow long-term studies. Nor are the benefits potentially provided by medical robots fully
understood. Medical robots have only passed through a few technological generations and the technology
continues to change and leap into new areas. Yet by looking at the current market and representative
research systems, educated guesses can be made about the impacts of robots on near-future medicine.

In surgical robotics, there has been a trend away from autonomous or even semiautonomous
motions, and toward synergistic manipulation and virtual fixtures. Thus, the robot acts as a guidance tool,
providing information (and possibly a physical nudge) to keep the surgeon on target. Such use requires
accurate localization of the tissues in the surgical site, even as the tissues are manipulated during surgery.
Improved imaging systems (e.g., Explorer, an intraoperative soft tissue tracker by Pathfinder
Therapeutics) or robot compatibility with MRI or CT will provide that localization. In particular, MRI-
guided robots will benefit from intraoperative 3D images with excellent soft tissue contrast and accurate
registration between the tool and the tissue, thus allowing precise virtual fixtures, “snap-to” and “stand-
off” behaviors. Further, such imaging will allow modeling and rapid prototyping of patient-specific
templates/jigs/implants.
 The physical designs for medical robots will continue to improve, reducing expense and size,
while minimizing or compensating for nonidealities such as flexion, for example, the CRIGOS robot.
With better physical designs, semiautonomous behavior will likely become more useful. “Macros” may
become commonplace, wherein the surgeon presses a button and the robot performs a preprogrammed
motion, such as passing a suture needle between graspers, or the Sensei’s autoretract feature.

Robots will see more use for medical training purposes, bolstered by improved tissue modeling
capabilities, by the increasing objectivity in healthcare assessment, by advances in computer simulations,
and as a result of increased data mining arising naturally from improved dataconnectivity between
devices and between institutions. Some such systems are already available, such as the aforementioned da
Vinci Skills Simulator, the Virtual I.V. Simulator by Laerdal, and the EndoscopyVR Surgical Simulator
by CAE. For the same reasons, robotics will continue to make possible new medical procedures and
treatments, such as new Single-Port Access procedures.

Even as robots are developed for new medical areas, other tools may encroach on medical needs
currently filled by robots. Medical robots must develop a firm basis in improved medical outcomes, or
risk being displaced by pharmaceuticals, tissue engineering, gene therapy, and rapid innovation in manual
tools (e.g., the SPIDER Surgical System by TransEnterix, and the EndoStitch by Covidien). To that end,
improvements in medical robotics must address and solve real problems in healthcare, ultimately
providing a clear improvement in quality of life when compared with the alternatives.
 2. Robotic Neurosurgery

2.1 Developments of robotic neurosurgery

Neurosurgical robotics had a long gestation period spanning over two decades. The main
reason for this long period of development is the stringent regulation of health and safety. In
contrast, industrial robots leaped into production very quickly because they can be isolated from
human contact in a cage or a highly secure environment; neurosurgical robots on the other hand
are designed to interact with surgeons and perform or assist the surgeon to perform complex
surgical procedures on alive but anaesthetised patients. Hence, the evolution of neurosurgical
robotics was slow as follows.

• The Unimation PUMA 200 (Advances Research & Robotics, Oxford, CT):

A standard industrial robot (PUMA 200) was used to hold a stereotactic biopsy needl in a
52-year-old man on a CT scanner table, the target was identified on the CT images and the robot
was used to orient a guide tube through which a needle was inserted (Kwoh et al., 1985).
Localization of the target was achieved by using the Brown-Roberts-Wells (BRW) stereotactic
frame localization plates and the head was secured to the CT scanner table using the stereotactic
frame reference ring. It is a programmable, computer-controlled, versatile robot that was
designed to perform highly accurate, delicate work, yet it was rigid enough to provide stable
trajectory. It was a safe robot, designed to work with humans and its joints were equipped with
spring-applied, solenoid-released brakes that automatically clamped should any mechanical or
electrical defect occur. It has 6 degrees of freedom; movements are executed by DC
servomotors; tracking is achieved by optical encoders and it can be used in passive or active
programmable modes. It has an accuracy of 2mm and repeatability of 0.05mm. It uses the
Brown-Roberts-Wells stereotactic frame for registration and CT scan for imaging. The use of the
cumbersome stereotactic frame is a constraint and as such its accuracy and performance are
similar to the frame, it has an advantage over the frame in those tedious calculations and manual
adjustments were automatically executed by the robot It was used as a retractor during resection
of thalamic astrocytomas (Drake et al., 1991) (Figure 1).
 Figure 1. The PUMA Robot (Courtesy Helge Ritter, Bielefeld University, Germany)

• The Minerva System (University of Lausanne, Switzerland):

The Minerva system was designed to perform within 5 degrees of freedom. It had two
linear axes (vertical and lateral), two rotary axes (moving in a horizontal and vertical planes),
and a linear axis (to move the tool to and from the patient’s head). The robot is mounted on a
horizontal carrier which moves on rails. A stereotactic frame, the Brown-Roberts-Wells (BRW)
reference frame, is attached to the robot gantry and coupled to the motorized CT table by two
ball-and-socket joints arranged in series. The system was used for two operations on patients in
September 1993 at the CHUV Hospital in Switzerland, but the project has since been
discontinued. The problems with this project were the limited degrees of freedom, the robot was
unwieldy and located within the CT scanner making the environment not ideal for performing
neurosurgical procedures and diagnostic imaging. It did not get rid of the cumbersome
stereotactic frame and as such it did not offer performance advantage compared to the frame. It
was fixed to the scanner making the procedure longer and was not cost effective as the CT scan
suite was unusable for other diagnostic scans during the procedure.

• Evolution 1 (Universal Robotics Systems, Schwerin, Germany):

 This robot was designed for both brain and spinal applications and has 6 degrees of
freedom. It is a hexapod robot based on parallel actuator configuration to provide a high degree
of accuracy and high payload capacity for drilling applications such, as drilling in

the spinal pedicles, and more laterally was used to steer a neuroendoscope (Zimmermann et al.,
2002).

• An MRI compatible robot (Masamune et al., 1995, Chenzie & Miller, 2001, DiMaio et al.,
2006):

This robotic system was devolped by Harvard Medical School in collaboration with
Mechanical Engineering Laboratory, AIST, MITI (Tsukuba, Japan). It has 5 degrees of freedom
and is MRI compatible. It works with intraoperative MRI system (Signa SP/1, General Electric,
Milwaukee, WI) and it has non-magnetic ultrasonic motors based on parallel configuration. It
consists of a three-degree-of-freedom Cartesian positioning stage and a two-degree-of-freedom
orienting mechanism, and is mounted above the surgeon's head in the open MRI magnet. Two
long rigid arms reach into the surgical space and form a parallel linkage for manipulating an
acrylic needle holder or guide. The five motion stages are driven by ultrasonic motors (Shinsei
USR-60N) attached to lead screws, and motion is measured by optical encoders with 10μm
resolution (Encoder Technology, Cottonwood, AZ). A flashpoint sensor is attached to the needle
holder to provide independent redundant encoding. This robot has been integrated with a
software planning interface (built into the 3D Slicer), and a tracking and control system for
percutaneous interventions in the prostate under MR-guidance. The surgeon interacts with the
planning interface in order to specify a set of desired needle trajectories, based on anatomical
structures and lesions observed in the patient's MR images. All image-space coordinates are
computed and used to automatically position the needle guide, thus avoiding the limitations of
the traditional fixed template guide. Once the needle holder is in position, the robot remains
stationary while the surgeon manually inserts the needle through the guide and into the tissue,
with real-time imaging for monitoring progress. The disadvantage of this device is its
dependence on intraoperative MRI scan and MRI compatible instruments. Whilst it is beyond
the reach of most centres worldwide today, it may become part of MRI technology in the future
as more and more surgery is performed at the time of diagnosis.

 The NeuroMate Robot (Integrated Surgical Systems, Davis, California, USA):

 It is commercially available and FDA approved and evolved from the work of Benabid’s
group in Gernoble University, France. It has 6 degrees of freedom, incorporates CT, MRI and
angiographic neuroimages. It was used in conjunction with a stereotactic frame to position a
cannula or probe for biopsy or targeting deep brain structures. It is a six-axis robot for
neurosurgical applications. The original system was subsequently redesigned to fulfil specific
stereotactic requirements and particular attention was paid to safety issues.(Figure 2).

Figure 2. The NeuroMate robot

To carry out a procedure by the NeuroMate, the robot must know where it is located
relative to the patient’s anatomy. This is typically done using a calibration cage, which is placed
on the end-effecter of the robot around the patient’s head. This cage looks like an open cubic box
and the four sides are each implanted with nine X-ray opaque beads, the positions of which have
been precisely measured. Two X-rays are taken which show the position of these beads along
with the fiducial markers of the patient’s frame. In the newer versions of this robot, an
ultrasonic-based registration is performed using the reference markers shown in Figure 3. This
information is used to determine the transformation matrix between the robot and the patient.
The defined trajectory is used to command the robot to position a mechanical guide, which is
aligned with this trajectory. The robot is then fixed in this position and the physician uses this
guide to introduce the surgical tool such as a drill, probe or electrode (Figure 3).

Figure 3. The NeuroMate robot during registration (courtesy TRK Varma, Liverpool)

• The CyberKnife (Accuracy Inc, Sunnyvale, CA):

It was designed for frameless stereotactic radiosurgery and its accuracy compares well to
localization errors in contemporary frame-based systems. The unique targeting capability of the
CyberKnife’s multi-jointed robotic arm uses a guidance system to track the location of tumours
in real-time and automatically adjusts its focus to a patient’s respirations to deliver high-level
radiation with pinpoint accuracy. This enables access to previously unreachable tumours with
faster, safer, and more comfortable treatments. The CyberKnife is an example of a robotic
system delivering treatment that the surgeon cannot do (Figure 4).
 Figure 4. CyberKnife radiosurgical system

The CyberKnife radiosurgical system is being used as a minimally invasive alternative to

traditional surgery in a variety of clinical areas in neurosurgery as well as other disciplines.

It offers an effective treatment option for patients who cannot undergo traditional open
surgery or whose lesions are inaccessible with traditional surgical approaches. Residual tumours
left after partial resection may also be treated. It has also been used as a boost to standard
radiation therapy and to treat failed surgery or radiotherapy. For intracranial conditions, the
CyberKnife system has been used to radiosurgically treat a variety of tumours such as residual
small skull base menigiomas, small acoustic schwanomas small pituitary adenomas, and small
metastases as well as other abnormalities such as small arteriovenous malformations (AVMs)
and intractable pain such as in Trigeminal Neuralgia (Massaudi et al., 2005). With the
Synchrony™ motion tracking system, tumours in organs moving with respiration such as the
lung (Brown et al., 2005), the pancreas (Goodman & Koong, 2005), the liver and the kidney can
be successfully targeted. Other tumours based in more rigid body anatomy, where minimal
motion is expected, may be tracked via rigidly implanted markers including those in the spine
and the prostate (Medbery et al., 2005). The CyberKnife system’s range of applications is
limited only by the imagination of clinicians who currently have, or will eventually have access
to this technology. To date, more than 10,000 patients have benefited from the revolutionary
concept of marrying robotics to image-guided radiosurgery.

Scientific presentations and publications on the clinical applications of the CyberKnife

are numerous – including intracranial

• The RoboSim Neurosurgery Simulator

This robotic neurosurgical simulator consists of a workstation and a robotic arm

(NeuRobot). The MRI image data-set is transferred into the system and the surgical target, its
coordinates and planning trajectories are programmed. It was developed as part of the
Roboscope project for minimally invasive neurosurgical procedures. Minimally invasive
neurosurgery is mainly of importance for treatment of diseases in the central area of the brain,
which is accessible to the surgeon only by transgression of healthy normal brain tissue, such as
hydrocephalus due to cystic brain tumors and ventricular tumours.

As we enter the 21st century, real-time simulation of surgical procedures is becoming the
norm in neurosurgical practice. The RoboSim is a robotic platform for surgical simulation and
planning minimally invasive and complex neurosurgical procedures. Another important aspect of
neurosurgery is the training of junior surgeons on how to anatomically orient them while
operating within the miniaturised operating field of minimally invasive procedures. Image-
guided simulation of the procedure will then allow the control of accessibility of the diseased
area along the pre-planned trajectory.

• The neuroArm :

The neuroArm is an MRI-compatible, ambidextrous robot. Its dextrous components are

two image-guided manipulators with end-effectors that mimic human hands and are capable of
interfacing with new microsurgical tools. It has tremor filters that eliminate unwanted hand
tremors seen under the microscope. It consists of a surgeon-machine interface and multiple
surgical displays. The interface consists of two hand controllers which hold tools. It has 8
degrees of freedom for each arm, payload of 0.5 Kg, a force of 10 N, tip-speed of 0.5- 5mm/sec
and submillimetric positional accuracy. It has optical and force sensors and can work
continuously for more than 10 hours.

• The PathFinder (Prosurgics, UK):

 The PathFinder is a neurorobotic system that is portable with a very stable base which
can be wheeled in and out of the operating room. The robotic arm can rotate in a horizontal plane
90 degrees to the left or right. The base fixes to the surgical space by an attachment to the
Mayfield head clamp. The proximal arm articulates with the next arm that moves in a vertical
plane that articulates with the third arm which again moves in a vertical plane. The most distal
arm holds the end-effecter, which can rotate 360 degrees and flexes/extends by 180 degrees. The
combined movements at all these joints give the PathFinder 6 degrees of freedom. The
PathFinder differs from other neurosurgical robots in that it does not require X-rays, ultrasound
or mechanical means to locate the surgical field; instead it depends on identifying reflectors
attached to the patient’s head using a camera system integrated in its head (Figure 5).

Figure 5. The PathFinder neurosurgical robotic system

The robot is driven by Windows® based task program and planning software. The
planning software and PathFinder robot detect the fiducial markers automatically with a
maximum accepted registration error of 1.25 mm. The tool holder is attached to the PathFinder’s
endeffecter. The predefined path is used to command the PathFinder to align its instrument
holder to the planned trajectory. Once the instrument holder is aligned to the trajectory, the robot
locks in position and instruments can then be passed to the predetermined depth such as probes,
electrodes or catheters (Figure 6).
 Figure 6. Tool holder holder of the PathFinder

• SmartAssist® (Mazor Surgical Technologies, Caesarea, Israel):

This miniature robotic system was designed to overcome the need to rigidly immobilise
the surgical field during robotic application. This robot achieved this by fixing the robot directly
to the bony element of the surgical field. This concept was clinically used in spinal pedicle screw
fixation using the SpineAssist robotic system (Shoham et al., 2007). The system consists of the
miniature robot that aligns the end effectors with 6 degrees of freedom and a workstation that
runs graphic user interface software and performs image manipulation, planning, registration,
kinematic calculations and real-time robot control. Once the system was assembled and
intraoperative registration using intraoperative fluoroscopy was performed, the plan for each
pedicle screw is executed by the robot and the surgeon manually drills the pilot drill-hole and
passes K-wire in the desired position. The SpineAssist is an automated pointing robot that gives
the surgeon full control.(1)

2.2 The Next Step in Robotic Neurosurgery

The idea of using a robot for neurosurgery seems like a (pardon the pun) no-brainer. At European
Robotics Week, currently being held in Brussels, the EU-funded Robocast demonstrated that it’s closer
than ever to making robotic neurosurgery a reality.
 No matter how steady a surgeon’s hands are, they’re still about 10 times shakier than Robocast’s
machine, which was designed by a group of scientists from Israel, the U.K., Germany and Italy. That
makes it perfect for delicate “keyhole surgery,” or surgery done through a tiny hole drilled through a
patient’s skull called a “burr hole.”

Needles and catheters inserted through the skull can then be moved by a probe operated by a
surgeon, who can feel the resistance of the brain thanks to an advanced haptic device which gives tactile
feedback to the operator of the machine. In the future, the probe will be used to perform surgical
procedures not possible today, such as allowing surgeons to take a curved path from an entry point in the
skull to a targeted lesion.

Ultimately, according to Robocast, the goal is to make “treatment quicker, less invasive, and
more effective,” especially when it comes to treating tumors, hydrocephalus, Parkinson’s disease,
Tourette syndrome and many other neurological diseases.

Of course, surgical robots are nothing new. They’ve been popular for treating prostate cancer for
years now (although not everyone agrees on their effectiveness). What’s so special about Robocast is the
machine’s level of autonomy.

Basically, the robot factors in the surgeon’s inputs along with other factors such as diagnostic
information from the patient as well as data gathered from optical, electromagnetic and ultrasound
sensors in the robot. The machine then proposes the most efficient, risk-free path for the procedure,
which is then accepted or rejected by the surgeon.

Robocast also has a higher standard for accuracy than most surgical robots, which, considering
the margin for error in brain surgery, makes sense. That would explain why, according to a European
Commission press release, it’s capable of performing 13 different types of movement at the operating
table, nine more than humans can do.

Right now, procedures have only been performed on dummies in a hospital operating room in
Verona, Italy and it’s perfectly understandable why it might be a while longer before we see this
technology used on human beings.

The real problem, however, isn’t the safety or efficacy of robots like the one developed by
Robocast; it’s the cost. Most surgical robots cost anywhere between $1-$2 million and require 150-200
procedures to master, according to the New England Journal of Medicine. That doesn’t even count other
expenses such as maintenance and replacing single-use appliances.

For relatively simple procedures, many wonder if surgical robots are worth the cost, especially
when you factor in the worry that robotically assisted surgery isn’t any more safe in the long-term than
normal laparoscopic surgery. While some of that is offset by the savings associated with shorter hospital
stays, hospitals still need to know that they’re going to use a robot enough to justify its cost.

So, will Robocast’s robot be worth it in the eyes of hospital administrators? It’s hard to say, but it
certainly seems to take a big step in the field of surgical robotics with its haptic feedback and (eventual)
curving probe. When you start talking about a robot performing procedures that no human has ever done
before, as opposed to simply assisting a human with a procedure, then the game arguably changes. That’s
when robotics’ relationship to medicine is altered forever.(8)
 3. NEUMESY Neurosurorgical Robot

3.1. The initial experiment description

In the last years a large number of new surgical devices have been developed so as to improve the
operation outcomes and reduce the patient’s trauma. Nevertheless the dexterity and accuracy required in
positioning the surgical devices are often unreachable if the surgeons are not assisted by a suitable
system. From a kinematic point of view, the robot must reach any target position in the patient’s body
being less invasive as possible with respect to the surgeon’s workspace. In order to meet such
requirements a suitable design of the robot kinematics is needed. This paper presents the kinematic design
of a special robot for neurosurgical operations, named NEUMESY (NEUrosurgical MEchatronic
SYstem).

NEUMESY is a six joints serial manipulator whose kinematic structure lets the robot to adapt to
different patient’s positions while minimizing the overall dimensions. Owing to the usual symmetry of a
surgical tool, the kinematic dimension of the neurosurgical task is five, being given by one point and one
direction on the space. Therefore the NEUMESY is kinematically redundant, leaving an extra DOF to the
surgeon to choose a suitable robot configuration which minimally limits his movements during the
surgical operations. The link lengths have been optimized in order to maximize the robot workspace with
respect to the surgical task, while minimizing the links static deformations.

The initial experimentation of robotic systems in surgery was undertaken during the early 1980s,
and it basically consisted of adapting the industrial robot technologies already in existence. In the last
decade, there has been a growing awareness, within the medical community, of the benefits offered by
using robots in various surgical tasks. Traditional surgery involves making large incisions to access the
part of a patient’s body that needs to be operated on. Minimally Invasive Surgery (MIS), on the other
hand, is a cost-effective alternative to open surgery. Basically, the same operations are performed using
instruments designed to enter the body cavity through several tiny incisions, rather than a single large
one. By eliminating large incisions, trauma to the body, post-operative pain, and the length of hospital
stay are significantly reduced.

However, new problems connected to the use of robots in surgery have arisen, since there is no
direct contact with the patient. For this reason, it is necessary to develop suitable tactile sensors to
provide surgeons with the perception of directly operating on the patient. Such a result can be achieved
by using force feedback systems, in which the force applied to patient’s tissue is fed back to a robotic
device (haptic master) directly operated by the surgeon.
 We can categorize surgical robots based on their different roles during surgical treatment Passive
robots only serve as a tool-holding device once directed to the desired position. Semi-active devices
perform the operation under direct human control. Active devices are under computer control and
automatically perform certain interventions. Moreover, the surgical robot can have different levels of
autonomy. To be specific, systems that are able to perform fully automated procedures are called
autonomous. On the other hand, when the surgeon completely controls every single motion of the robot,
this is called a tele-operated system or a master-slave system. Medical robotics has found fruitful ground
especially in neurosurgical applications, owing to the accuracy required by the high functional density of
the central nervous system.

In past decades, several different robotic neurosurgical devices have been created. A
comprehensive survey can be found in. In particular, in the 1980s Benabid and colleagues experimented
with an early precursor to the robot marketed as Neuromate (Renishaw Mayfield, UK). Today’s robot
projects focus on three major areas of improvement: increasing the overall accuracy of the classical
stereotactic systems increasing the added-value of the equipment enhancing the capabilities of the
surgeon

The Mechatronics Research Group (composed of researches of the University of Padova,

University of Udine and University of Trieste, Italy) with the assistance of the Neurosurgical Department
of the University of Florence has developed two master-slave robotic systems for minimally-invasive
neurosurgical operations. The first robot (Figure 7), named LANS (Linear Actuator for Neuro Surgery)
has been conceived specifically to perform biopsies and neurosurgical interventions by means of a
miniaturized x-ray source (the PRS, Photon Radiosurgery System, by Carl Zeiss), whose emitting tip
must be placed accurately inside the patient’s brain tissues .

Figure.7 LANS LANS (Linear Actuator for Neuro Surgery)

 The LANS robotic system is composed of a haptic master module, operated by the surgeon, and
a slave mechatronic module moving a PRS probe, or a biopsy needle, along a predefined emission axis in
accordance with the master position imposed by the surgeon. In order to orient the LANS along the
established emission axis, a NeuroMate robot is employed in a frame-based configuration which ensures
the highest possible accuracy. The system has been designed assuming that during the surgical operation
only the LANS (which is very accurate, and provides the surgeon with force feedback) is in active mode
while the NeuroMate is powered off. This allows overcoming much of the problems associated with the
complex nature of this surgical therapy. Moreover, very precise and repeatable movements of the biopsy
needle and of the x-ray source can be obtained, thus improving the overall intervention outcomes.

DAANS (Double Action Actuator for NeuroSurgery) is the second robot (Figure 8) carried out
by Mechatronics research group. The aim of the system is to provide another degree of freedom to the
PRS source about the emission axis.

Figure 8. Simplified kinematic structure of the DAANS

The system allows extending the therapy with PRS also to irregular shape tumorous lesions, by
integrating translation and spin movements of the source. Indeed, single isocenter radiosurgery
procedures produce nearly spherical isodose distributions. In order to avoid unacceptable dose delivery to
non target tissues, the high-dose region must be shaped to fit individual targets. The high-dose region can
be manipulated into a variety of shapes that closely conform to a tumour shape by means of shielding
caps and multiple targets dose delivery. The caps are placed on the probe of the x-ray source as in Figure
7. Nevertheless, LANS and DAANS limit the NeuroMate mobility, owing to their geometrical
dimension, which can interfere with the robot arm movements (Figure 9). In this manner the NeuroMate
workspace is reduced and some tool configuration is not reachable . This paper presents the results of a
preliminary analysis on the kinematic structure of a new surgical robot, named NEUMESY. The robot is
able to maximize the performances of the therapy by means of the DAANS and PRS, increasing the
workspace of the overall robotic system and allowing all possible tool configurations on the patient’s
head to be reached.

Figure 9. The Neuromate and DAANS Systems

3.2 Medical and Robotics requirements

For sterilization reasons no non-medical equipment must be closer than a fixed distance from the
operating site. Thus the robot must collide neither with the patient, nor with the medical staff, nor with
the surgical tools.

The requirement becomes quite complex in the frame-based applications, where the stereotactic
frame (the mechanical structure the patient’s head is fixed on) interferes with the robot movements. In
order to satisfy this requirement, it is useful to define a virtual sphere (Figure 10) including the patient’s
head and the stereotactic frame. The robot must neither cross nor touch this sphere. The radius of the
sphere depends on the stereotactic frame dimensions, while the center is on the patient’s head. Moreover,
the robot must minimally limit the surgeon’s movements during the operation. To this end, the surgeon
must be able to choose a suitable robot configuration for each tool position and the robot must adapt its
workspace to the surgeon’s requirements, which may change during the surgical task.
 Figure 10 . The virtual sphere includes the patient’s head and the stereotactic frame

The robotics constraints concern the structure of the NEUMESY in order to satisfy the
above medical constraints. The preliminary choice considers the robot kinematic structure. The
advantages of a serial robot if compared to a parallel one are due to the larger workspace and the
higher dexterity and manipulability. On the other hand a parallel manipulator is stiffer allowing
higher accuracy in the tool positioning. According to the medical constraints, the robot has to be
able to avoid the virtual sphere and it must minimally limit the surgeon’s movements. Therefore,
the solution adopted consists in a serial structure.

Nevertheless this choice requires a links length optimization so as to maximize both the
workspace and the stiffness of the robot.

Moreover, since the neurosurgical tools have usually an axial symmetry, only two spatial
points on the patient’s head have to be stated by the surgeons. The first one is the Target Point
(TP), the center of the cerebral lesion where the tool has to be placed, while the second one is the
Entry Point (EP), the hole through which the surgical instruments go into the skull. EP and TP
state the Line of Action (LoA) along which the tools should be moved (Figure 11). Since the
surgical operation fixes only five constraints on the space, a five DoF manipulator should be
sufficient to a neurosurgical operation, nevertheless, a further DOF yields the kinematic
redundancy which allows infinite configuration for the same surgical task, letting the surgeon
choose the suitable one.
 Figure11. NEUMESY, the DAANS and the virtual sphere 3 robot configurations for the same
TP and LoA

3.3 Robot kinematics

The structure of the designed robot and its Denavit-Hartenberg description are displayed in
Figure 10 and in Table 1, respectively. The first prismatic joint allows the robot to adapt to the patient’s
vertical position and change the robot configuration according to the surgeon requirements. The next
three revolute joints form a spherical wrist whose position reduces the load and the elastic displacements
on the robot links. Finally, two revolute joints allow orienting the end-effector while keeping the robot
outside of the virtual sphere.
 Table.1 Parameters of the NEUMESY

The robot features six DoFs while the kinematic dimension of the neurosurgical task is five,
being given by one point and one direction on the space. Therefore the NEUMESY is kinematically
redundant. As stated above, the redundancy can be used by the surgeon to choose a suitable robot
configuration which minimally interferes with his movements. Indeed, the origins O2-O6 belong the the
same plane on the sheaf (of planes) defined by the LoA. The plane is arbitrary but fixes the robot
configuration and the constraints on the surgeon’s movements. By choosing a different plane, the space
required by the robot for the surgical operation changes and the surgeon’s movement could be easier. The
solution of the robot inverse kinematic will be presented in the next section.

Solution of the inverse kinematic

For sake of simplicity, let Ti be the matrix i−1 Ti , as well as R i =i−1 Ri the rotation matrix from
i−1
the frame i to the frame i −1 and pi = pi the origin of the frame i referred to the frame i −1. The
inverse kinematic problem can be expressed as
(
0 T 6
= T 6 T −1
tool
=: T = s n a p (1)
0
( ( ( ( 0 0 1
(s n a p
where T =
0
0 0 1
Equation (1) can be rearranged as
2 (2)
T6 = Q1
where:
2
T6 = T3 T4 T5 T6 and Q1 = (0 T2 )−1 T
Equation (2) can be rewritten as:

T3 T4 T5 T 6 = (0 T2 )−1 T ( A 6 )−1 (3)

1 0 0 a6
R6 0 0 0 −1 0
where: T6 =T and A =
0 1 6 0 1 0 0
0 0 0 1
Comparing to each other the left and the right sides of (3), the following equations must hold:
 a=R3R4R5R6z
(4)
p = R 3 R 4 p 5 + R 3p4
where a and p are defined by: ( 0 T2 ) −1 T ( A 6 ) −1 =
s n a p
and z =[0, 0,1]
T
. Equations (4)
0
0 0 1
can be further transformed into:
R −41 R 3−1 a = R 5 z
R −41 R 3−1p = p 5 + R 4−1 (5
)

(6
After multiplying left and right sides to each other of the equations (5), it is possible to write: )

p·a = p 5 ·( R 5 z ) +(R 4−1p 2 )·(R 5 z)

where the right side of (6) is zero, since the vectors therein involved are orthogonal. Therefore from

(6) the value of the q1 can be determined, since the projection of p into a depends only on this joint

(( ( (((

d t a ⋅ n + a6 s ⋅ n − p ⋅ n
(
variable. In particular: q =− −d q. When n = 0 the z -axis belongs to the
1 ( 2 z 0

nz

operation plane. In this case, if the target point belongs to robot workspace, there are infinite solutions for q1 .
Taking into account the module of p , there is:

p·p = p 4 ·p 4 + p 5 ·p 5 + 2 (R 4 p 5 )·p4 (7)

the left side of (7) being noted, since the module of p depends only on joint variable q1 . In this way

the value of q5 can be determined from (7):

p·p −l 22 −l23
sin q5 = (8)
2l 2 l3

The equation (1) can be rearranged as

1
T = Q , where 1
T = T T T T and Q = T −1 T T−1 .
5 2 5 2 3 4 5 2 1 6

Considering the new vectors:

%
 s=R2R3R4R5x (9)
%
p=R2R3R4p4 + R 2 R 3p 4 +p2
−1 −1
% % % %
T
where a and p ar e def ined by: T1 TT6
% % s n a p
= and x =[1, 0, 0], rearranging (9) as

0 0 0 1
−1 %
( R2 R 3 ) s= R4R5x
( R 2 R 3 ) −1 %
( p −p 2 )= R 4 p5 +p4
and operating as in (5) it is possible to write:
% % −1 (10)
(p − p 2 )·s = p 5 · ( R 5 x ) + (R 4 p 4 )·(R 5 x)

The left side of (10) depends only on joint variable q6 , while the right side is known, once q1
and q5 have been calculated. In this way (10) can be expressed in the form: Ac6 + Bs6 = C , where
( (( ((
A = a6 s ·s + d t s ·a − s ·p + sz q1
(( (( (( (
B = a6 s ·a + d t s ·s − s ·p −az q1

C = − ( l 3 +l2 s5 )
The solutions are: q = atan2(± A 2 + B 2 − C 2 ,C) −atan2(B , A) .
6
1
Equation (1) can be rearranged as: T 5 = Q3
l
3
I
where 1 T5 = T T T T5 , Q = T −1 T T −1 A−1 , R5 0 3 0
2 3 4 3 1 6 5 T 5= T and A5 =
0 1 0
T
0 1
Considering the vectors:
aˆ = R 2 R 3 R 4 R 5 z (11)
pˆ = R 2 R 3p 4 +p2
where ˆ and ˆ are defined by Q3 = sˆ nˆ aˆ pˆ and rearranging as above, it is possible to write:
a p

0 0 0 1
R 2− 1aˆ = R 3 R 4 R 5 z (12)
R 2−1 (pˆ −p2 )= R 3p4 (13)
From the projection of the equation (13) along y =[0,1, 0]T , the value of c can be determined
− d + pˆ ( ( ( ( ( 3
as c = 2 z where pˆz = ( − a z ) d t + p z − q1 −a z l 3 s6 − a6 s z −c6 l3s z
3
l2
while from the projection of the equation (13) along z, it is possible to write pˆy c2 = pˆ x s2 where
( ( (
pˆ x = p − a x(d t + l 3 s ) − ( a +c l 3 )s
(x ( 6 6 6 x

=p −a
 (
pˆ (d + l s ) − ( a +c l )s
y y y t 3 6 6 63 y

In this way: q2 = atan2( pˆ x , pˆ y ) + k π with k = 0,1

Finally, projecting (12) along z, the value c4 is determined as: c4 = aˆ x s2 −aˆ y c2 , where
of
( (
aˆ =n and aˆ = n . Therefore: q = atan2( c , ± 1−c2 ) . There are four solutions for each pose
x x y y 4 4 4
defined by the matrix T , which mainly differ to each other on the elbow configuration. During the
surgical operation only the DAANS is in active mode while the NEUMESY is powered off. Therefore the
requirement on the accuracy on the tool positioning concerns only the robot in the static configuration.
At the same time, for safety reasons, the contacts between the robot and the virtual sphere have to be
avoided.
The link lengths affect both the accuracy and the robot workspace owing to the NEUMESY kinematic
structure and have to be appropriately chosen in order to satisfy the requirements stated above. The
requirement on the robot workspace can be described by a reachabilty index which gives information
about the number of points on the patient's skull being achievable along each desired tool
configuration.
desired tool orientations at
Pi bedescribed by all the vectors zk
internal to the cone Λα (Pi ) withvertex at Pi , angle α and axis
coincident with the normal to the skull surface in Pi (Figure 12).
Moreover, let the function IK(P , z) calculate the number of solutions
for the inverse kinematic problem:
P = P
tool
z
tool = z
subject to the constraint that all the robot links are external to the
virtual sphere. The reachabilty index can be defined as:
N
1
Figure 12. Cone Λα
Φ
l2 , l3
=
∑Θα ( Pi ), Pi ∈ S
N i=1
where the function Θα returns 1 if, for each vector zk belonging to the cone Λα ( Pi ) , the function
IK( P , z) is nonzero. The function Θα returns 0 otherwise. The index Φl 2 , l3 achieves the maximum
(i.e. 1) when every point Pi ∈S is reachable along any direction internal to its cone Λα ( Pi ) .
In order to maximize the robot workspace, the skull surface is taken coincident with the virtual
sphere. Therefore the generic point can be described as Pi = R[cos ϕi cos θ i ,cos ϕ i sin θ i ,sin ϕi ]T

where R is the radius of the virtual sphere and ϕi and θi are the polar angles of the point. The angle α
is chosen coherently with the tool orientation allowed during a neurosurgical task. A suitable value
is α = 70° .
.9
.9
4
0
.
98
0
.98 Table 2. Parameters
employed in the
0
.92 0

700 0 0
.

optimization problem
98
96 .

600
0 0
.
9 0 . 1
. 94
92
500 0 0 .
.

96 98
L
3

R
1
400 0. 1
9 ..
0 0
92 94

300 0. 0. 0
.98

0. 96 98
5 0 .
.96
0
9 0. .94 0
.92 0
.9
200 94 0
200 300 400 500 600 700
L2

Figure 13: Level curves of Φl 2 , l3 when the robot parameters

of Table 2 are assumed
Figure 13 shows the level curves for Φl 2 , l3 as a function of the link lengths l2 and l3 for the robot
parameters in Table 2. It can be noticed that the index value is maximum inside the region R, while
outside of this regionΦ l 2 , l3 <1, which means that some tool configurations are not reachable by the
robot. Since both the index Φl 2 , l3 and the arm stiffness depend on the link lengths, a suitable
optimization problem must be defined so as to maximize the robot workspace while keeping the
position errors, due to the links deformations, to a minimum.
The optimization problem can be stated as:
max 1
3
l 2 , l3 >0 ( l 2 +l3 )
Φ =1
s.t. l 2 , l3

P ∈ S % l 2 , l3 > 0
where the cube at the denominator of the objective function takes into account the fact that the link
stiffness at the end point is inversely proportional to the cube of link length. With reference to
the values of Table 2, the maximum is reached at l2 = 660 mm and l3 = 355 mm .

In order to highlight the performances of the NEUMESY for the chosen l2 and l3 values, let the function
Θα be modified in Θα ( Pi ), so that it returns not only a boolean value (0 or 1) but the mean
number of solutions for each tool configuration defined by

Pi ∈ S and z k ∈ Λα (Pi ) :
Θα ( Pi )= 1
M
∑ IK ( Pi , zk )
z k ∈ Λα ( Pi )
k =1..M

30

Figure 14 shows the Θα ( Pi ) function for −2

20 −1

the values in Table 2. It can be noticed that 0

1 x
2 1
3
there are on average more than ten admissible
z

10
Figure 14. Performance index Θα ( Pi )}
solutions for each tool configuration.
0

In order to have a clearer idea of the robot 0

performances, it is convenient to give a spatial

 4
5
representation of the Θα ( Pi ) index, by y 6
7
2

introducing a suitable surface SΘα defined as:

3 1
SΘ = {x ∈ |x= Θ (P)P,
α
P ∈ S}
α R
where R is the radius of S.
20 20 25

15 15 20

10 10 15

5 5 10

0 0 5

5 5 0
z

y
10 10 5

15 15 10

20 20 15

25 25 20

30 30 25
30 20 10 0 10 20 30 15 10 5 0 5 10 15 15 10 5 0 5 10 15
y x x

Figure14. Projections of SΘ on the cartesian planes

α

Figure 15 shows the projections of the SΘ on the three principal planes. It can be noticed that
α

the highest performances are achieved on the lateral sides of the virtual sphere, where there is the
largest number of configurations allowing the robot to avoid the virtual sphere.

In this paper a new robot for neurosurgery is presented. The robot, named NEUMESY, is a six joints
serial manipulator designed as positioning device for a neurosurgical actuator previously built by our
research group, able to move a low energy x-ray source for the treatment of cerebral lesions. The
redundancy respect to the surgical task (5 DoFs, being given by one point and one direction on the
space) gives an extra DoF allowing the surgeon to choose the suitable robot configuration which
reduces the space required by the robot and maximizes the safety, keeping the robot links away from
the patient's head.

The solution of the non-trivial inverse kinematic problem is produced and a kinematic simulator has
been carried out in order to analyze the performances of the robot.

Finally, the links length has been optimized in order to satisfy the workspace requirements fixed by
the neurosurgical task and reduce the static deformations of the robot arm.

Future works consist in the study of the trajectory planning using the benefits taken by the redundancy
so as to limit the link vibrations and deformations during the surgical task.
3.4 Conclusions

The robot, named NEUMESY, is a six joints serial manipulator designed as positioning
device for a neurosurgical actuator previously built by our research group, able to move a low energy
x-ray source for the treatment of cerebral lesions. The redundancy respect to the surgical task (5
DoFs, being given by one point and one direction on the space) gives an extra DoF allowing the
surgeon to choose the suitable robot configuration which reduces the space required by the robot and
maximizes the safety, keeping the robot links away from the patient's head.

The solution of the non-trivial inverse kinematic problem is produced and a kinematic simulator has
been carried out in order to analyze the performances of the robot.

Finally, the links length has been optimized in order to satisfy the workspace requirements fixed by
the neurosurgical task and reduce the static deformations of the robot arm.

Future works consist in the study of the trajectory planning using the benefits taken by the redundancy
so as to limit the link vibrations and deformations during the surgical task. (5)
 References

1. Robotic Applications in Neurosurgery M. Sam Eljamel

2. Robotic Neurosurgery 175 Karas, CS and Baig, MN

3. Efficient Design of Precision Medical Robotics

4. The Application Accuracy of the NeuroMate Robot— A Quantitative Comparison

with Frameless and Frame-Based Surgical Localization Systems Qing Hang Li, M.D.,
Ph.D

5. NEUMESY: A special robot for neurosurgery Paolo Boscariol

6. http://www.hindawi.com/journals/jr/2012/401613/

7. http://www.renishaw.com/en/neuromate-stereotactic-robot-
10712#ElementMediaList55684

8. http://techland.time.com/2011/11/30/the-next-step-in-robotic-neurosurgery/