Robotics in medicine

zyxw zyxwvuts zyxwvut Robotics in Medicine Paolo Dario, Eugenio Guglielmelli, Benedetto Allotta ARTS Lab Scuola Superiore Sant'Anna via Carducci 40,56127 Pisa, Italy Abstract zyxwvutsrqpo zyxwvu zyxwvuts zyxwvutsrq This paper reports the current state-of-the-artin medical robotics. Rather than trying to enumerate all the possible applications of robots or robotic technologies to medicine, three general areas of advanced robotics are identified which, based on the current technological background and expertise, could potentially provide significant improvements to the state-of-the-artin medicine. These are: mucro roboiics, micro robotics and bio-robotics. Macro robotics includes the development of robots, wheelchairs, manipulatorsfor rehabilitation as well as new more powerful tools and techniques for surgery; micro robotics couM greatly contribute to the field of minimally invasive surgery as well as to the development of a new generation of miniaturised mechatronic tools for conventional surgery; bio-robotics deals with the problems of modelling and simulating biological systems in order to provide a better understanding of human physiology. According to this classification, a review on the most important past and ongoing research projects in the field is reported. Some commercial products already appeared on the market are also mentioned and a brief analysis of the economical potentialities of robotics in medicine which can be prefigured in the near future is presented. 1. Introduction The application of robots - or more generally of technologies and know-how derived from robotics research to medicine has moved rapidly in the last few years from the speculation of a small group of "visionary" scientists to reality. Today robotics can be considered as a real opportunity, available to a range of operators in the medical field, as well as to industries which want to explore a market that can quickly become very attractive [ 11. The growth of interest on medical application of robotics has been so rapid recently, that is already difficult to provide a "still" picture of this field. However, we propose a classification that may be helpful as a guideline to discuss the main applications or perspectives of robotics in medicine. This classification is presented in Figure 1. Rather than trying to enumerate all the applications of robots or robotic technologies to medicine, we have identified three main areas of advanced robotics which, based on the current technological background and expertise, could potentially provide significant improvements to the state-ofthe-art in medicine. Robots can find practical application in two main medical fields: surgery and assistance to disabled and elderly persons. Moreover, robotics can also be conceptually associated to biological systems in the area that we can broadly define as bio-robotics. In this area, robotics can be seen as a "metaphor" of biological systems, and robotics research as an important bridge between human and biological sciences, on one side, and artificial intelligence, on the other side. 739 Figure 1 - Main applicationsand perspectives of medical robotics However, instead of being just speculative, research on bio-robotics may also lead to a number of practical applications for the substitution or augmentation of organs andtor functions of humans. zyxwvutsrqpo zyxwvutsrq zyxwvu A distinction that can be helpful to clarify some basic concepts and to discuss applications of medical robotics is the one we propose between "macro robotics" and "micro robotics". In fact, this distinction implies not only different size, but - and most important - different intrinsic design features and mode of operation. This is particularly clear in the area of surgery, which most scientists and industrialists perceive as the most promising field of application for robotics, and where two substantially different approaches can be envisaged: the use of robots as rigid accurate and autonomous machine tools, on one hand, or the use of robotlike endoscopes, flexible and teleoperated by the surgeons, on the other. Somewhere in the "middle" is an area also very interesting, which some investigators refer to as "mechatronic tools for surgery". The aim of this branch of medical robotics is to broaden the concept of robotic device for surgery by taking advantage of methodologies and technologies directly derived from the state-of-the-art in robotics. Considering the valuable progresses of the last decades in the field of micromechanics and mechatronics, this approach could potentially lead to a quick and wide "diffusion" into the market of innovative surgical tools and thus to clinical practice. Significant potentialities have been also identified in the field of rehabilitation robotics. For instance, the feasibility of desktop robotic assistants has been already demonstrated, even though under particular conditions [Z,31. In spite of this, the potential of mobile robotic systems has not been clearly defined yet. In the last decade, the prototypes of robots dedicated to household tasks which have been commercialised did not encountered the favour of the market, mainly due to their high cost associated to poor performance with respect to the average expectation of the user. In addition, the market for rehabilitation robotics is far from being a single one, actually being still strongly dependent on the trends of the mass consumer market. However, the success of such systems will mainly rely on their modularity and easiness of use, which must be considered as key factors from the very beginning of system design. In fact, modular components and friendly user interfaces could represent the real link between the mass consumer market and the rehabilitation technology market. Assistance systems currently available on the market require heavy adaptation of the house by means of special building design, installation of centralised environmental control systems (for door, windows, appliances, etc.) and of fixed desktop workstations. These systems are rather expensive, especially if also the cost of building special apartments and residences is taken into account. On the contrary, mobile personal robots represent a highly attractive solution, even in economical terms, as they could significantly contribute to minimise the required degree of adaptation of the house. This fact will not only decrease the global cost for the installation of the assistance system but it will also have some functional and cultural implications: the house will be available for use to both able-bodied and disabled people, no specialised environment will be necessary apart from the rest of the house, and consequently no problems will arise if the disabled person needs to move to a new apartment. In this paper, an overview of the past and ongoing research projects as well as of the first commercial products already appeared on the market is reported for the areas of surgical and rehabilitation robotics. Finally, basic concepts and potential applications of bio-robotics will also be discussed with reference to current research activities in the field. 2. Robotics for surgery Many different projects in this field have been carried out during the last ten years and few of them already generated industrial systems that are currently under clinical evaluation in hospitals [4,5,6]. The numerous applications to surgery can be classified in two main areas: those based on "image-guidance'' and those aimed to obtaining minimal invasiveness. zyxwvutsrq 740 2.1 Image-guided surgery The basic concept behind image-guided surgery is the use of a robot workstation integrated into the operating theatre where some of the parts of the patient's body are fixed by means of suitable fittings. This scenario is easy to implement for orthopaedic surgery, where fixators are commonly used to fix bones and also for neurosurgery, where the stereotactic helmet, mounted on the patient's head, is quite popular to provide absolute matching between pre-operative and intraoperative reference frames. Vision-based surgery may be viewed as a robotic CADCAM system where diagnostic images (from CT, NMR, US, etc.) are used for off-line planning of the intervention. The robot is used in a CNC machine tool-like fashion for precise cutting, milling, drilling and other similar tasks. A better quality of the intervention results from better performance of robots with respect to the manual operation of a surgeon in terms of accuracy and repeatability. A clear demonstration of the superiority of robot cutting versus normal cutting is shown in Figure 2 for the case of bone milling for hip implant PI. Figure2- A comparison between (a) robot's and (b) surgeon's performance in bone milling for hip replacement [SI. Real-time images may also be used during the intervention in combination with diagnostic images and tool positiodorientation data in order to provide the surgeon with feedback about the current state of the intervention. It is important to point out that the surgeon supervises the robot system during operation. Among the obvious differences between an industrial robot application and a surgical one, the need for suitable matching procedures between diagnostic images and off-line intervention planning on one hand and real execution on the other hand is still a key issue. As mentioned before, the issue of matching has been addressed and solved in some cases (specifically in the case of bone cutting in orthopaedics), but many problems still remain open due to the fact that most interventions on parts of the human body involve soft tissues and large deformations may occur. This results in possible discrepancies between pre-operative and intra-operative images. Image-guided surgery includes orthopaedic surgery, spine surgery, neuro-surgery, reconstructive/plastic surgery and ORL surgery. A very representative example of implementation of image-guided robotic surgery is the one proposed by R. H. Taylor et al. [6] which has been implemented in an industrial system (Robodoc, Orthodoc, ISS Inc., Sacramento, CA, USA) currently used in human trials for the automated implant of hip prostheses. zyx zyxwv The main reason which motivated many researchers to explore the use of robotic devices to augment a surgeon's ability to perform geometrically precise tasks is the consideration that the precision of surgical planning often greatly exceeds that of surgical execution. The ultimate goal of this effort is a partnership between men (surgeons) and machines (computers and robots) that seek to exploit the capabilities of both to perform a task better than either can perform alone [5]. The architecture of the hip replacement surgery system depicted in Figure 3, consists of a CT-based pre-surgical planning component, shown in Figure 4, and of a surgical component. The surgical procedure includes manual guiding to approximate positions of pins pre-operatively inserted into bones (which are f i x a d to the operating bed) and automatic tactile search for each pin. Then, the robot controller computes the appropriate transformation between CT and robot coordinates and uses this information to machine out the implant cavity. Finally, the pins are removed and the surgery proceeds in the conventional manual procedure. zyxwvutsrqp zyxwvutsrq zyxwvutsr Figure 4 - An example of the pn-operative planning procedure for the hip replacement system [6] Safety issues have been taken into high consideration. In the "Robodoc" system they include extensive checking [7] and monitoring of cutter force, and the possibility for either the surgeon or the controller to freeze all robot motion or to turn off manipulation and cutter power in response to recognized exceptions. Techniques which are essentially similar to the one described before, but which have been adopted to different operation tasks and scenarios, have been developed for the cases of total knee arthoplasty [8] (see Figure 5), percutaneous discectomy [9],spine surgery [IO] (see Figure 61, neurosurgery [ l l , 121, prostate surgery [13], and eye surgery (by the group of Ian Hunter at the McGill University, Canada). zyxw Figure 3 - A view of the hip replacement surgery system in the operational theatre [ 5 ] 74 1 Similar approaches have been presented by many authors, for example by Finlay and Giorgi 1141 for neurosurgery (see Figure 7), by Stuttem er d.for om surgery 1151. zyx zyx zyxwvu zyxwvutsrqpon zyxw zyxw Figure7- Advanced man-machine interfaces and force replication devices might also play an important role in the framework of intervention simulation and surgeon training carried out in virtual environments featuring realistic 3D representations of body organs. Some examples of interfaces possessing the sophisticated features which are required for truly realistic simulations of surgical interventions are already existing, like the one developed by Bergamasco et al. 1161 at the Scuola Superiore Sant'Anna (Pisa, Italy) and many more will probably appear in the near future, like the one under development at the McGill University (Canada) by Hunter et al.. (c) Figure 5 - An example of knee prosthesis implant : (a) before the intervention; (b) after the intervention; (c) detail of femur resection obtained by robot cutting [8] 2.2. (a) Figure 6 - A stereotactic helmet equipped with passive arm for neurosurgery [ 141. Minimal invasive surgery Minimal invasive surgery (MIS), also called "endoscopic surgery", is gaining increased acceptance as a powerful technique beneficial to patient's integrity time of recovery and cost for assistance. At its current stage of development, MIS depends on three prerequisites: the availability of high quality video endoscopy, the ability of high precision surgical instruments and the manual skill of well-trained surgeons ~71. MIS requires accessing the organ to be operated through a small hole, and the surgeon, although directly responsible for the manipulation of the surgical tool, misses large part of the information necessary to control finely the end effector. At present MIS is a sort craftmanship, where operating surgeons has to compensate with their skills the fact that they can not touch and sense with their fingers for diagnostic purposes, they may not have 3D view of the workspace, the access to the workspace is restricted, they can not feel the forcedtorques and pressure they are exerting at the end effector tip. A possible scenario of the next generation of MIS consists of a combination of telemanipulation and telerobotics technology, as depicted in Figure 8. (b) An example (a) of computer assisted spine surgery for treating a case of scoliosis (b) [IO] Teleoperation, virtual reality environments and advanced madmachine interfaces will probably play a key role in the future of image-guided surgery. Teleoperation can be useful in some cases, such as when a patient must be operated urgently in a place where no specialised surgeon is available (for example, on a battle field or an ambulance), or when for safety reasons (patients with infective diseases, or long operations under X-ray) it is not appropriate for the surgeon to be within the operation field. Another critical problem which can be solved by means of teleoperation is the unavoidable increasing requirements in terms of room for equipment in the neighbourhood of the patient's bed, so that it might become necessary for the surgeon to move away and operate remotely. 742 zyxwv A scheme of the system for brain blood vessels diagnosis and surgery is depicted in Figure 10. As outlined by Sato et al. [ 191 and by Morishita et al. [20] (see also Figure l l ) , the development of suitable instrumentation for MIS requires considerable research efforts both in miniaturisation of components and the adaptation of teleoperation techniques [ 191. zyxwvutsrqponm zyxwvutsrqpon Figure 8 - A possible scenario for teleoperated surgery [ 171 Generally speaking, the contribution of robotics to significant decrease of the level of surgical invasiveness could involve three different fields: the first is laparoscopic surgery, which is based on stiff tools that the surgeon manipulates directly and by which can keep some (even if small) degree of "sensation" on the features of the operation workspace; the second is commonly referred as endoscopic surgery which makes use of flexible endoscopes and implies the virtual loss of any type of "sensation" for the surgeon; the third is not linked to any specific type of surgery and consists of an attempt of improving the performances of traditional macro surgical tools by applying mechatronic technologies aimed at decreasing the invasiveness of tool operation. Major developments efforts are needed in such areas of robotics technology as sensor integration, force reflection, miniaturisation of mechanisms and actuators, control. zyxwvutsr zyxwvutsrq zyxwvuts 4Yu Figure 10- Scheme of the system for brain blood vessels diagnosis and surgery [18j. zyxwvutsrqpon zyxwvutsrq Figure 9 - Example of catheter tip with increased dexterity [IS] A very challenging approach to MIS is pursued in Japan, where the development of teleoperated micro-catheters capable of diagnostic and surgical interventions within brain blood vessels is currently underway. The micro-catheter will possess high dexterity at the tip like the macro one shown in Figure 9, and all along its length and will incorporate microfabricated tactile flow and pressure sensors at the tip, along with micronozzles and micropumps for local injection of drugs and solutions for dissolving thrombus [18]. -information disphcement, s o d and SO 011 displrcement, R e d d m of displacement, force and dimensionsof tools Figure 11 - Relations between micro world and human world [20]. 743 zyxwvutsrqp An example of next generation micro endoscopes for MIS is given in Figure 12. Numerous research teams, potential users and manufactures are already involved in developing techniques for environmental control, and for controlling robots in the context of rehabilitation and assistance for disabled users. At present, one of the commonest use of rehabilitation technology puts a general purpose computer at the hub of a multi-purpose 'cockpit', and the users operate all of their (specialised or adapted) products from that cockpit [22]. There are some advantages to this, especially for severely disabled or bed-ridden users, but less so for users with moderate, or age-related disabilities: the user is distanced from the task itself, the interaction style is based on the computer, rather than on the product being used and the task being performed. In such a context of use, and for these users, the computer remains an obtrusively technical device which tends to appear as the unique link between disabled users and their environment. zyxwvutsrqpon zyxwvutsr zyxwvuts 3.1 Figure 12 - A next generation micro endoscope for MIS [18]. 3. Robotics for rehabilitation The possible role of robotics in the field of rehabilitation has been widely investigated in the last decades. Possible specific applicative areas which have been already identified range from the assistance to the disabled and the elderly, by means of robotic manipulators, intelligent wheelchairs and dedicated interfaces for household and vocational devices, through the restoration of impaired functions, by means of advanced prostheses, ortheses and electrical functional stimulation (FES), to the development of virtual environments for training and genuine rehabilitative therapies. Whatever is the selected approach, one of the key factors for the success of robotic aids is certainly the potentiality to make these peculiar users still able to exert a complete control on their environment by using a robotic interface. In rehabilitation robotics, the term "environmental control" refers to a disabled user's capacity to actively interact with his or her external environment [21]. Although all of the sensory and motor functions are necessary for a complete environmental control, disabilities based on partial or total loss of upper limb function are particularly serious, due to the consequent reduction in, or loss of, the manipulative function. This kind of disability is the most significant impediment in carrying out common everyday activities (e.g. personal hygiene, job, hobbies): the user receives the external stimuli but is then unable to respond to, or act on them (by modifying the external environment, for example). When lower limb function is also reduced (or lost), the physical (and psychological) loss of control is profound, and makes a disabled user dependent on others in virtually every respect. 744 Manipulators in rehabilitation One of the primary objectives of rehabilitation robotics has always been to fully or partly restore the disabled user's manipulative function by placing a robot arm between the user and the environment. Some important factors must be considered in the design of such a peculiar environmental control system: the user's degree of disability (a system must be flexible enough to be adapted to each user's capabilities); modularity (system inputs and outputs must be easy to add or remove according to each user's needs); reliability (a system must not let the user down); and cost (the system must be affordable). According to the state-of-the-art of rehabilitation robotics, three different configurations of robot systems, differently reflecting the above mentioned factors, have been considered as feasible for the assistance of severely and moderately disabled users. Historically, the first configuration which has been investigated is the bench- or table-mounted manipulator included in a completely structured desktop workstation [23, 24, 25, 26, 27, 281. Even though various systems based on this approach were positively evaluated with users [2, 3, 28, 29, 301 and some commercial products already appeared on the market, such as the DEVAR system ("olfa Corporation, Palo Alto, CA, USA) and the RAID system (Oxford Intelligent Machines Ltd., Oxford, Great Britain), yet they seem to be particularly suitable for assisting disabled employees for executing vocational tasks at their workplace. In fact, desktop workstations better reflect the type of organisation of space and time which is typical of vocational activities. Furthermore, this approach brings all the previously discussed drawbacks of using a 'cockpit' environmental control system. One of the first prototypes of desktop workstation, the MASTER (Manipulator Autonomous at Service of Tetraplegia for Environment and Rehabilitation) system was developed in France by CEA (Paris) and is shown in Figure 13. zyxwvuts zyxwvut zyxwvutsrq zyxwvutsrq zyxwvutsrqp The wheel-chair-mounted arm is particularly suited to users with upper limb disabilities, but its usefulness relies on the user being able to move (or control the movements of) the wheelchair, so that the robot can be taken to the area(s) of activity: the home environment must also be adapted to suit the robot's working height, making it an intrusive solution. This solution, depicted in Figure 14, is becoming much popular as it allows to the disabled or elderly to use the robot arm anywhere, not being necessarily related anymore to some fixed structured locations [31, 321 and is being widely experimented [33,34]. A first prototype of a mobile robot for rehabilitative applications was developed by S. Tachi et al. of the MITI Japanese laboratories. This system, named MELDOG,was devised only to act as a robotic "dog" for blind patients, thus not having any possibility of manipulating or carrying objects. zyxwvuts zyxwvutsrqpo Figure 1 4 Figure 13.- The MASTER workstation (CEA -France) However, some technical problems, mainly concerning the accuracy which can be obtained on grasping operations (the arm is fixed to the wheel-chair which is not properly a rigid structure) and the possibility to equip the wheel-chair with a friendly (and then complex) arm controller (the available space and the battery energy are limited) have still to be solved. Moreover, this solution does not actually address all disabled users, but only those with upper limb disabilities. Consequently, it seems to have limited concrete perspectives to become economically attractive by inducing mass market demand. In both previous cases, a severely disabled or bed-ridden user is not well catered for: the workstation dominates the user's home environment, while the wheelchair-mounted robot is simply not an option. A third different solution is that of using an autonomous or semi-autonomous mobile vehicle equipped with a manipulator and additional sensor systems for autonomous or semi-autonomous operation. Its mobility and versatility make it particularly suitable for severely disabled or bed-ridden users, as long as the interface between the user and the robot is easy to use: the user should be able to instruct the robot with a high-level language, via a bi-directional user interface offering appropriate methods of input and suitable output. This configuration has been first used in industrial applications (e.g. textile industries) and is surely the most sophisticated one, but it is also the most generally applicable, since the idea to have a robot in personal service can result attractive for both able-bodied and disabled users. 745 - The wheelchair-mountedMANUS manipulator [32]. Generally speaking, the ideal system should include subsystems dedicated to vision, fine manipulation, motion, sensory data acquisition, and system control. A sketch of a possible configuration for such a robot [l] is shown in Figure 15. Unfortunately, the cost of this solution is often prohibitive for all but the wealthiest of users, and the usability problems inherent in such sophisticated systems have yet to be satisfactorily solved for non-professional users. Various prototypes of autonomous or teleoperated mobile robots for the assistance to the disabled in different activities have been implemented and others are currently under development [35, 36,371. Figure 15 - Concept of a mobile robotic aid [I] As a sample, Figures 16 and 17 illustrate two ongoing research projects in this field: namely, the american MOVAR system, mainly devised for vocational use, and the Italian URMAD system, mainly devised for household applications. Both prototypes have been mostly implemented. In particular, the URMAD system (the acronym stands for "Mobile Robotic Unit for the Assistance to the Disabled") will be widely experimented with the final end users, i.e severely disabled patients, before the end of 1994. Figure 18, the final objective of this project, in fact, is to develop a complete system, including a mobile robot, having globally functional performances similar to URMAD but still respecting the limitations imposed by a normal household environment so to avoid heavy adaptation of the house. zyxwvutsr zyxwvutsrq zyxwvutsrqp I I zy -1 LIVINoRom I Figure 18 - The MOVAID approach a mobile base which fits into different activity workstations. This goal is being pursued by partially distributing resources in the environment instead of concentrating them on the vehicle. As long as distribution of resources is well balanced and technologies are properly integrated in the domestic environment, such a realistic solution could represent a good compromise between the current state-ofthe-art in advanced robotics and the ideal concept of the autonomous robotic assistant in a modern domestic scenario, thus hopefully favouring a rapid commercial exploitation. To this aim, the MOVAID project, rather than developing new basic hardware components, will take advantage of the results of previous research projects. As a sample, in Figure 19 the URMAD manipulator, which will be also used for the MOVAID mobile unit, is shown. zyxwvutsrqpon Figure 16 - The Mobile Vocational Assistant (MOVAR) [I]. One of the main unsolved technical problems is that the more the performance of the robot is enhanced the more its dimensions (imposed by the standard building norms), its autonomy of operation, and consequently its cost become excessive for prefiguring an effective commercial exploitation. -_-F MOBILE BASE zyxwvutsrqpon Figure 17 - An overview of the URMAD system An attempt [38] to overcome these drawbacks and limitations is being carried out in the framework of the TIDEMOVAID project by an European team coordinated by the ARTS Lab of the Scuola Superiore S. Anna. As illustrated in Figure 19 - The URMAD manipulator: an 8 d.0.f. redundant dextrous arm purposely developed for service applications (SM - Scienzia Machinale srl - Pisa, Italy). 746 zyxw zyxwvutsr zyxwvutsrqpo Another important aspect of the MOVAID project is that the emphasis is on the friendliness of the system's interface and on a non-intrusive integration of technologies in the domestic environment. In this optics, MOVAID represents a potential opportunity to have a direct transfer of advanced technologies towards the home environment. This approach, which is strictly related to the increasing industrial interest in "domotics" (i.e. the development of a "smart" house accessible to all users, including the disabled and the elderly) could also have interesting implications in the medical field (telemedicine, home assistance vs. clinical assistance for the disabled and the elderly, etc). 3.2 Robotic systems for hospitals Mobile robots could become one answer to the current shortage of help in hospitals, as well as one solution to the problem of the diseases (lumbago, low back pain) which affect the personnel involved in heavy physical tasks such as lifting a patient and carrying himher to the toilet or changing sheets in the bed. An example of implementation of a hospital transport mobile robot has been presented recently by Transition Research Corporation (TRC Inc., Danbury, CT, USA). The robot, depicted in Figure 20, has been designed and built for addressing the need for assistance with such tasks as point to point delivery [39]. estimation, and moving obstacles (e.g. people). HelpMates have been installed in several hospitals. In some hospitals HelpMate is in operation 24 hours per day and the hospitals are reporting an increase in productivity and efficiency. The HelpMate represents a useful and probably industrially valuable solution to some basic needs requiring the transportation of lightweight objects. The complexity of the system is deliberately kept low by eliminating automated manipulation (which is carried out by human operators), by assuming flat floor in the working environment (the robot uses elevators - which are controlled by means of elevator control computers activated by an infrared transceiver-, and doors), and by providing the robot with accurate geometric and topographical information about the hospital hallways, elevator lobbies and elevators. Further help to nurses could be provided by heavier robots, designed to execute tasks requiring hard muscular work. Japanese laboratories and industries have identified this field as very promising, and have invested substantial efforts in the development of fetch and carry robots for hospitals, usually hydraulically actuated and featured by high payload. An interesting example of such type of robot is the patient care robot named "MELKONG", that was developed a few years ago by the Mechanical Engineering Laboratory (MEL) in Japan [40].The "MELKONG was intended to lift, hold and carry an adult patient (weighing up to about 100 kg) or a disabled child. The robot docked to the bed in the hospital room, lifted the patient in its arms from the bed, moved back still holding himher in its arms and transferred himher to the toilet, or bathroom, or dining room. Usually the robot was controlled by a nurse, but it was expected that at night the patient could also call the robot and control it by means of simple commands given by means of a joystick. Serious problems related to automatic docking, mobility, manipulation, actuation (by hydraulic actuators), energy supply, madmachine interfaces were addressed and solved. zyxwvutsrqpon zyxwvutsrqpo Figure 20 - The HelpMate hospital transport mobile robot [39]. The objective of the hospital transport mobile robot ("HelpMate 0 " )is to carry out such tasks as the delivery of off-schedule meal trays, lab and pharmacy supplies and patient records. The navigation system of HelpMate, unlike many existing delivery systems in the industry which operate within a rigid network of wires buried or attached to the floor ("AGVs"), relies on sensor-based motion planning algorithms that specifically address the issue of navigation in a partially structured environment. The system is also able to handle sensor noise and sensor inaccuracy, errors in position zyxwvutsr 747 Figure 21 - Recent version of a robot for lifting bedridden patients [41] zyxwvu zyxw zyxwvutsrq zyxwvutsr zyxwvuts The MELKONG concept has evolved from the early prototype described above to more sophisticated versions incorporating functional and aesthetic improvements, such as the one depicted in Figure 21. A transfer-carrier vehicle based on an evolution of the MELKONG concept is commercially available in Japan from Sanyo Electric Co. Ltd. Furthermore a simple functional robot aimed at supporting the elderly and the disabled for independent living, in particular for evacuation (a function that the elderly wished to do by himherself if an adequate support equipment is available), is also being developed in Japan jointly by the National Institute of Bioscience and Human Technology of AIST, Aprica Kassai Inc. and Hitachi Ltd, in the framework of the National Programme for Welfare State and Apparatus (Towards a New Society). 3.3 Intelligent wheelchairs For a physically disabled person the main advantage of using a transport mechanism, like a wheelchair, is to allow himher to achieve some degree of personal mobility. In the case of a wheelchair carrying a robot arm, the severely disabled can use the robot arm anywhere, and not only remotely or in fixed permanent locations. In order to increase the performance of ordinary wheelchairs (and of the robot manipulators possibly mounted on them), a number of approaches have been proposed based on robotic and mechatronic technologies. These approaches comprise attempts to develop autonomous vehicles which can be used to transport a person from one location to another with little or even without outside assistance, as well as attempts to increase the capability of the vehicle to locomote on unprepared surfaces and to overcome obstacles. An example of the first approach is represented by selfnavigating wheelchairs, as the one proposed by Madarasz et al. a few years ago [42]. The vehicle, designed to function inside an office building, is able to plan its own path from its current location to a particular room in the building, and then to travel to that location. The system must also function with minimum impact on the building in which it will be used, that is the building cannot be equipped with a guidance mechanism, such as embedded wires in the floor or painted stripes that can be followed. Therefore, the wheelchair becomes substantially a sort of mobile robot with high degree of autonomy. In fact, the vehicle is self-contained: all of the sensing and decision making are performed by the on-board equipment. This approach relieves the disabled person from tasks heher may be unable to carry on, but a system for supervised control is provided for high level commands and for other types of operations requiring direct guidance. In Europe, a recent example of a project aimed at developing a wheelchair featured by partly autonomous behaviour is represented by the European TIDE-OMNI project. An italian manufacturer (TGR s.r.l,, Ozzono Emilia, Italy) 748 produces a wheelchair (named "Explorer") incorporating both wheels and tracks. This wheelchair, which has been also modified to host the Manus system in the framework of the European SPRINT-IMMEDIATE project, can not only run on regular terrain, but also go up and down stairs with the user on board. This approach represents an evolution towards the possible development of a new generation of vehicles designed to deeply enhance the mobility of the user. A very interesting example of this evolution is the adaptive mobility system proposed recently by Wellman er al. [43] at the University of Pennsylvania, Philadelphia, USA The design of the system, that is basically a hybrid vehicle incorporating wheels as well as two "arms" that can work both as manipulators and as legs, is based on the assumption that a legged vehicle allows locomotion in environments cluttered with obstacles where wheeled or tracked vehicles can not be used. A legged vehicle is inherently omnidirectional, provides superior mobility in difficult terrain or soil conditions (sand, clay, gravel, rocks, etc.) and provides an active suspension. The legs also give the chair versatility and allowed it to be re-configured. When stationary, one of the legs can be used as a manipulator in order to perform simple tasks such as reaching for objects or pushing open doors. 4 From advanced prostheses and ortheses to F.E.S. There have been frequent intersections between robotics and limb prosthetic technologies in the past. Many devices, like artificial legs, artificial hands and arms, have evolved in the '60s and '70s both as prostheses for amputees and as possible components of advanced robots. Examples of these devices are the Belgrade hand 1441 and the UTAH arm [45]. More recently the UTAH-MIT dextrous hand [46] was designed as a robotic hand by taking inspiration from the human hand, whereas new prostheses for amputees have been developed by exploiting last advances in robotic technology (like the one developed in the framework of the European TIDE-MARCUS project). However, it is quite obvious and very clear to all of those working in the field of aids for disabled that, although disabled persons may accept "artificial" devices as assistants, their dream and ultimate goal is to be able to manipulate and walk again. Although this is out of reach for current medical capabilities, a few promising approaches are being pursued by some investigators which might ultimately lead to render that dream closer to reality. A first example of robotic device that has been developed to help patients with impaired walking capabilities to restore their functions is the one developed in Japan and illustrated in Figure 22. An evolution of the above mentioned assistive device for "natural" walking is the active orthosis, whose development has been pioneered by Prof. Pierre Rabishong and his team of INSERM, in Montpellier, France [47]. This active orthosis, zyxwvutsr zyxwv which is shown in Figure 23, was intended and used for clinical research and re-education. A further application of the system was the combination of the re-education apparatus with functional electric stimulation (F.E.S.) studies. This evolution is the goal of a EUREKA project, called CALXES Computer-Aided Locomotion by Implanted ElectroStimulation), which represents probably the most important coordinated effort presently carried out in the world for restoring autonomous locomotion in paralysed persons. The project investigates the possibility of implanting stimulating electrodes into lower limb muscles, or even nerves, and to obtain close to natural walking by providing appropriate stimulation patterns by means of an extemal portable computer. In fact, some projects aim to develop implantable neural prostheses based on the capability of the peripheral nervous system to regenerate, which could establish a bidirectional electrical continuity between the nervous system and extemal devices [49, 501. One of these projects, the European ESPRIT-INTER project [5 11, investigates an approach based on a combination of silicon microfabrication technology, polymer channel guidance and nerve growth factors release, in order to promote nerve regeneration through a pattern of microholes with electrodes, thus rendering it possible to pick up sensory signals and to selectively stimulate nervous fibers. Other projects are exploring the possibility of by-passing interruptions in the nervous system (even at the central level) by artificial nerve grafts, fabricated by conductive polymer fibers. Some other projects try to obtain nerve regeneration at the spinal cord level by implanting fetal nervous cells. It is the hope of many researchers that this class of "hybrid" devices, which exploit the properties of artificial materials combined with those of biological factors (possibly modified by means of biotechnology), could eventually allow to restore the continuity of nervous pathways, and make the dream of paralysed persons to move their own limbs true. 5. Bio-robotics zyxwvutsrqpon zyxwvutsrqpon zyxwvutsrq Figure 22 - Robotic device for walking rehabilitation [48]. A key component for improving the performance of FESbased apparatus is the implanted electrode. Development in this field might not only allow to obtain better systems for computer-assisted manipulation and locomotion, but eventually even lead to achieve the dream of restoring natural manipulation and locomotion. Figure 23 - Active orthosis: master-slave version [47]. 749 Biological systems are not only the recipient of the services of robots, but also the source of inspiration for components and behaviours of future robot systems. This not well defined , but intriguing and stimulating area of interest for robotics, can be called "bio-robotics" and is currently receiving an increasing amount of attention by many investigators. In general, biological systems are a living proof that some complex functions (both sensorimotor and "intellectual") that robotics researchers should like to realize in artificial systems can actually be implemented. Locomotion, manipulation, vision, touch are all functions which living beings execute seemingly without effort, but which turned out to be extremely difficult to replicate in artificial systems. In the recent past, many different groups have been active in this "borderline" area where the distinction between "robotics'hnd "bioengineering" becomes very subtle. A book discussing state-of-the-art results and perspectives in this field has been published recently 1521. Examples of components which are explicitly inspired to their biological counterparts (and which are intended to be the "core" of sensorimotor systems capable of replicating the function of their biological counterparts) are retina-like CCD .sensors [53] and tactile sensors [54]. A photograph of the CCD vision sensor, whose geometry is inspired to the one of the human retina (including the high-acuity fovea-like central part), is reported in Figure 24. A scheme of a fingertip incorporating three different types of sensors which provide (in combination with appropriate sensorimotor acts) the robot controller with information on object geometry and material features comparable to those of zyxwvuts zyxwvutsrq the human fingertip, is given in Figure 25. Further applications in the field of bio-robotics involve the use of artificial systems as accurate models for investigating the physiology of biological systems. The laboratory which has pioneered this approach is the one headed by the late Prof. Ichiro Kato at Waseda University, which has developed robotic devices capable of playing different musical instruments such as organs, piano, violin and flute. A system developed for investigating the function of mastication in humans is depicted in Figure 26. zyxwvutsrqpon zyxwvuts Figure 24 - The f o v d structlrre of the CCD retina-like sensor [53]. A close-up of one of the sensors (a 256-element array sensor), which imitates the space-variant distribution of tactile receptors in the fingertip skin, thus emphasising the role of "attentive behaviour'' in active touch, is also reported in the same figure. Figure 26 -The mastication robot [ S I . Prof. Kato himself (ICNR '91) and other investigators (Coiffet [%I, Rabischong [57]) have proposed even more intriguing speculations on the relations between human mind and robot mind. Based on these hypotheses the Waseda laboratory is currcntly investigating an approach to the assistance to the disabled and the elderly which involves not merely the concept of "service robots", but even the one of robot as "companions" of humans. 6. Conclusions (b) Figure 25 - The ARTS Lab fingertip probe (a) and tactile sensor (b) [%I. The reasons why robots did not gain immediate acceptance in the medical community are in part obvious. and in part more subtle. The obvious reasons are both psychological (robots may be perceived as "competitors" by physicians, and as potentially dangerous exotic machines by patients) and technical (industrial robots are reliable. but no real expertise exists in the world about robots working full time in the vicinity or even in contact with humans). The subtle reasons are related to a possible misconception of the very same notion of robots, which should probably be corrected in the interest of the robotic research community. In fact, most users perceive robots either as the industrial robot arm, or as an exotic and anthropomorphic creature. In the field of advanced robotics. and of medical robotics in particular, the robot leaves the factory floor and gets into physical contact with the human operator (the surgeon, the patient). In some cases, the robot will maintain the overall usual structure of an industrial robot (although new robot 750 arms dedicated to medical application are being presented), but in most other cases robotic technologies will be embedded into tools which will not possess the traditional robotic "look". This shift should not be seen, in our opinion, as a problem, but rather as a very interesting and attractive opportunity for the robotics research community to extend their reach to a broader area (sometimes referred to as "mechatronics in medicine" or even "bio-mechatronics"). Nevertheless, the concrete experimental and clinical results achieved both in the fields of "macro" and "micro" medical robotics, and only partially reported in this paper, together with the economical and social motivations for using these new technological tools, permeated by robotic technologies, could certainly represent the best viaticum for a massive development of this new area of advanced robotics in the near future. 121 N. VILLO7TE. D. GLAUSER, et al.: Conception of stereotactic instruments for the neurosurgical robot Minerva, Roc. of the 14th Annual Int. Conf. of the IEEE EMBS '92.1089-1090, Paris, France. 131 B. DAVIES: Soft tissue surgery, Proc. of I s f European. Conference on Medical Robotics (ROBOMED '94). June 20-22, 1994, pp. 67-72 141 C. GIORGI, M. LUZZARA. D. S . CASOLINO, AND E.ONGANIA:. A computer controlled stereotactic arm: virtual reaIity in neurosurgical procedures, in Advances in Stereotactic and functional neurosurgery edited by B. Meyerson, Springer Verlag, Vienna, NY, USA, 1993, pp. 75-76. [I51 J. SThTEM, B.KREMER, B. KORVES, L. KLIMEK, G. KRUCKELS, R. MOSGES. Robotics in ear, nose and throat surgery, Proc. of 1st European. Conference on Medical Robotics (ROBOMED '94).June 20-22.1994, pp. 31-35. [I61 M. BERGAMASCO, B. " I T A , L. BOSIO, LFERRETTI, G. PARRINI, G. M. PRISCO, F. SALSEDO, G. SARTINI: An arm exoscheleton system for teleoperation and virtual environments applications. Proc. of the 1994 IEEEInt. Conference on Robotics and Automation, San Diego, CA, USA, May 8-13,1994, pp. 1449-1455. [I71 E. HOLLER, W. WEBER: System and control concepts for a telemanipulator system to be applied in minimally invasive surgeq, Proc. of the I s r IARP Workshop on Micro Robotics and Systems, Karlsruhe, Germany, June 15-16, 1993, pp. 111-120. [ 181 H. NARUMIYA: Micromachine technology (Intraluminal diagnostic & theraptic system), Proc. of the 1993 Workshop on Micromachine Technologies and Systems, Tokyo, Japan, October 26-28, 1993, pp. 60-65. [19] T. SAT0 et al.. Micro handling robot for micromachine assembly, Proc. of the Isr IARP Workshop on Micro Robotics and Systems. Karlsruhe, Germany, June 15-16, 1993, pp. 138-146. [20] H. MORISHITA, Y. HATAMURA: Development of ultra precise manipulator system for future nanotechnology, Proc. of the 1st IARP Workshop on Micro Robotics and System, K a r l ~ ~ h Getmany, e, June 15-16, 1993, pp. 34-42. [21] L. LEIFER: Rehabilitative R o b e , Robotics Age, May/June, 1981. [22] G. BIRCH. Technology and Homes: the Remote Gateway Project, The Neil Squire FoundationJournal. Technical section, 5-8, 1992. [23] H. ROESLER, V. PAESLAK Medical Robot arms, Proc. of 1st CISM-IlToMM Symp. on Theory and Practice of Robots and Robof a m , Udine, Italy, 1974, pp. 158-169. [24] J. GUITTED et al.: The Spartacus, Telethesis: robot arm control studies, Bulletin of Prosthetics Research, 2, 1979, pp. 69-105. [25] H. FUNABUKO et al: Application of robot arm for environmental control system of bed ridden patients in private house, Proc. of I Ith Int. Symp. on IndustrialRobots, Tokyo, 1981, pp. 71-78. [26] M. R. HILLMAN. A feasibility study of a robot manipulator for the disabled, Journal of Medical Engineering & Technology, 1987, pp. 160-165. [27] D. LEE3 et al.: A third-generation desktop robotic assistant for the severely physically disabled, Proc. 1 Irh Annual Conference on Rehabilitation Technology (ICAART '88).Montreal, Canada, 1988. [28] W. SEAMONE, G. SCHMEISSER: Early clinical evaluation of a robot d w o r k t a b l e system for spinal-cord-injured persons. Journal Rehabilitation Research and Development, 22,L. 1985, pp. 37-56. [29] L. HOLMBERG, Rehabilitation Robotics at HADAR, Rehabilitation Robotics Newsletter, Vol. 4. No. 4, Wilmington, DE, Fall, 1992 [30] R. CAMMOUN, J. M. DETRICHE, F. LAUTURE, B. LESIGNE: Evaluation of the MASTER 1 robotic system and design of the new version, Proc. IEEE Int. Con$ on Advanced Robotics (ICAR '93), Tokyo, Japan, 1993, pp. 319-322. [31] S . D. PRIOR: A wheelchair mounted telemanipulator ann for use by the disabled, Research Project Report 1988, Middlesex Polytechnic, London. [32] H. KWEE, C. STANGER: The Manus robot arm, Rehabilitation Robotics Newsletter, 2,2, 1993, pp. 1-2. [33] J. C. ROSIER, J. A. VAN WOERDEN, L. W. VAN DER KOLK, B. J. F. DRIESSEN, H.H. KWEE, 1. J. DUIMEL, J. J. SMITS, A. A. TUINHOF DE MOED, G. HONDERD, P. M. BRUYN: Rehabilitation robotics: the MANUS concept. Proc. of '91 ICAR, vol. 1, pp. 893-898, Pisa, June 1991. zyxw zyxwv zyxwvutsrqponmlkjihgfed zyxwvutsrqpon zyxwvutsrqpo zyxwvutsrqpon zyxwvutsrqp 7. References J. F. ENGELBERGER: Robotics in service, Biddles Ltd., Guildford (GB), 1989. J. HAMMEL, K. HALL, D. LEES, L. J. LEIFER, H. F. M. VAN DER LOOS, 1. PERKASH, R. CRIGLER, Clinical evaluation of a desktop robotic assistant, Journal of Rehabilitation Research and Development,2,3, 1-16, 1989. J. HAMMEL, H. F. M. VAN DER LOOS H. F. M., I. PERKASH, Evaluation of a vocational robot with a quadriplegic employee, Archive of Phys. Med. and Rehabil., 2,683-693, 1992. P. CINQUIN, J. DEMONGEOT, J. TROCCAZ, S . L A V A U E , G. CHAMPLEBOUX, L. BRUNIE, F. LEITNER, P. SAUTOT, B. MAZIER, A. PfiREZ. M. DJAID, T. FORTIN, M. CHENIN, ANDC. A. IGOR: Image guided operating robot: methodology medical applications, results. ITEM (Innovation and Technology in Biology and Medicine) - Special Issue on Robotic Surgery, 13(4):373-393, 1992. H. A. PAUL,, B. MITIZESTADT, B. MUSITS, R. H. TAYLOR, P. KAZANZIDES, J. ZUHARS, B. WILLIAMSON AND W. HANSON: Development of a surgical robot for cementless total hip arthoplasty, Clinical Orthopaedics, December 1992, Vol. 285, pp. 57-66, USA. P. FLURY, P. LOPEZ, D. GLAUSER, N. VILLOll'E, C. W. BURCKHARDT: Minerva, a robot dedicated to neurosurgery operations, Proc. of the 23th ISIR, Barcelona, Spain, 1992, pp. 729733. R. H. TAYLOR, H. A. PAUL, P. KAZANZIDES, B. MITIZESTADT, W. HANSON, I. ZUHARS, B.WILLIAMSON, B. MUSITS, E. GLASSMAN, W. L. BARGAR: Taming the bull: safety in a precise surgical robot, Proc. of '91 ICAR. International Coneference on Advanced Robotics, vol. 1 , pp. 865-870, Pisa, June 1991 M. MARCACCI, M. FADDA, P. DARIO, S . MARTELLI, A. VISANI, T. WANG: The total knee arthoplasty project: a demonstration feasibility of the computer assisted approach, Proc. of 1st European. Conference on Medical Robotics (ROBOMED '94), in, June 20-22, 1994, pp. 59-66. M. FADDA, S . MARTELLI, P. DARIO, M. MARCACCI, S . ZAFFAGNINI, A. VISANI: First steps towards robot-assisted discectomy and arthroscopy, ITEM (Innovation and Technology in Biology and Medicine) - Special Issue on Robotic Surgery, 13(4):394-408,1992 P. SAUTOT, S . LAVALL6E. J. TROCCAZ, P. CINQUIN: Computer assisted spine surgery, Proc. of 1st European. Conference on Medical Robotics (ROBOMED '94). June 20-22, 1994. pp. 51-58. Y. S . KWOH, J. HOU, E. A. JONCKHEERE, S . HAYATI: A robot T guided with improved absolute positionning accuracy for C stereotactic brain surgery, IEEE Transactions on Biomedical Engineering. Vol. 35, N. 2, February 1988, pp. 153-160. zyxwvutsrqpo zyxwvutsr 751 zyxwvutsrqponmlk zyxwvutsrqp zyxwvutsrqpo zyxwvutsrqponm zyxwvutsrqp zyxwvutsrqpo zyxwvutsrq zyxwvutsrqp zyxwvut [34] J. HENNEQUIN, R. G. S. PLAITS: IEE Colloqium Digest. No.1992/108. [35] Regress Report, Veterans Administration R & D Center, 1988, Pa10 Alto CA, USA. [36] M. A. REGALBUTO, T. A. KROUSKOP, J. B. CHEATHAM: Toward a practical mobile robotic aid system for people with severe physical disabilities, Journul of Rehabilitation Research and Development, E, L 1992, pp. 19-26. [37] P. DARIO, E. GUGLELMELLI, C. MULE', M. DI NATALE: "URMAD: An Autonomous Mobile Robot System for the Assistance to the Disabled, 6th International Conference on Advanced Robotics ('93 ICAR), Tokyo, Japan, November 1-2.1993, pp. 341-346. [38] P. DARIO, E. GUGLIELMELLI: L'introduzione di nuove tecnologie nell'ambiente domestico. Atti del Convegno "Tecnologie.servizi ed edilizia innovativa per la terza etci.",SAIEDUE, Bologna, Italy, 1994. [39] J. EVANS et al.: Handling real-world motion planning: a hospital bansport robot, IEEE Control System Magazine, 1992, pp. 15-19. [40] E. NAKANO et al.: F'irst approach to the development of the patient care robot, Proc. of Ilth Int, Symp. on Industrial Robots, Tokyo, 1981, pp. 95-102. [41] S . HASHINO Aiding Robots. Advanced Robotics, 2.1,1991, pp. 97103 [42] R. L. MADARASZ et al.: The design of an autonomous vehicle for the disabled, IEEE J. of Robotics and Automation, Vol. RA-2, No. 3, 1986, pp. 117-126. [431 P. WELLMAN. An adaptative mobility system for the system. Proc. of the 1994 Int. Coni on Robotics and Automation, San Diego, CA, 1994,3. pp. 2006-201 I [U] G. BEKEY et al.: Control architecture for the Belgrade-USC hand, Proc. of 1988 IEEE Int. Con& on Robotics and Automation, Philadelfia, PA, USA, 1988. [45] S. C. Jacobsen et al.: Development of the UTAH artificial arm,IEEE Transactions on Biomedical Engineering, BME-29(4), pp. 249-269. [461 S. C. Jacobsen et al.: Design of the UTAWMIT dextrous hand, Proc. of the I986 IEEE Int. Conference on Robotics and Automation. San Francisco, CA, USA, pp. 1520-1532. 752 POI J. P. MICALEF'F AND P. RABISHONG: Motorized assistance for the handicapped: control systems, vehicles, walking machine, Proceedings of the First European Workshop on Evaluation of Assistive Devices for Paralyzed Persons, 1984 pp. 125-132, Edited by A. Pedotti and R. Andrich. COMAC BME. T. DOHI: Robot technology in medicine, Advanced Robofic.s,Vol. 7, NO. 2, pp. 179-187, 1993. D. J. EDELL: A peripheral nerve information transducer for amputees: long-term multichannel recordings from rabbit peripheral nerves, IEEE Transactions on Biomedical Engineering, Vol. BME-33, no. 2, February 1986. pp.203-213. G. T. A. KOVACS, J. M. ROSEN: Regeneration-type peripheral nerve interfaces for direct d m a c h i n e communication, in Robots and biological system.: towards a new bionics?, NATO AS1 Series, Springer Verlag, Berlin, 1993, Vol. 102, pp. 637-666. M. COCCO et al.: An Implantable Neural Connector lncorparating Microfabricated Components, Proceedings of MicroMechanics Europe, Neuchatel, Switzerland,, 1993, pp. 185-1 88 P.DARI0, G . SANDINI, P. AEBISCHER: Robots and biological systems: towards a new bionics?, NATO AS1 Series, Springer Verlag, Berlin, 1993. P. DARIO, M. DE MICHELI, G. SANDINI, M.TISTARELL1: Retinalike ccd sensor for active vision. NATO ARW on Rubots and Biological Systems, II Ciocco, Tuscany, Italy, 1989. Springer-Verlag. M. RUCCI, P. DARIO: Autonomous leaming of tactile-motor coordination in robotics, Proc. of the I994 IEEE International Conference on Robotics and Automation, San Diego, CA, May 8-13, pp. 3230-3236. H. TAKANOBU, A. TAKANISHI, I. KAT0 Design of a mastication robot mechanism using a human skull model, Proc. of '93 IROS, Yokohama, Japan, July 26-30,1993. pp. 203-208. P. RABISCHONG: Is m still the best robot, Proc. of Isr European. Conference on Medical Robotics (ROBOMED '94). June 20-22, 1994, pp. 157-159. P. COIFFEP Robot habilis and robot sapiens or some ideas about robot man links in robot design, Proc. of 1st European. Conference on Medical Robotics (ROBOMED '94),June 20-22, 1994, pp. 169174.