Copyright © IEEE 1995. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

Rehabilitation Robotics in Europe

John L Dallaway and Robin D Jackson
Department of Engineering
University of Cambridge, UK

Paul HA Timmers
Directorate General XIII
European Commission, Belgium

Abstract

This paper provides a summary of European work in Rehabilitation Robotics. The historical background to current work in this field is described. The European Union's TIDE funding initiative is introduced and recent projects established under this initiative are summarised. The practical effect of TIDE on the European market for rehabilitation equipment is discussed. European projects funded from other sources are also described. The paper concludes with an assessment of the future direction for research and development in Rehabilitation Robotics.

1 Introduction

The object of this paper is to provide a summary of European work in Rehabilitation Robotics with particular emphasis on the influence of the European Union's collaborative funding initiative, TIDE. Although the paper concentrates on European work, it is important to note that the development of equipment and practice in Rehabilitation Robotics cannot be considered in isolation from other work, particularly that in North America.

A convenient starting point is the pioneering work in France on the Spartacus project and concurrent work in Germany on a workstation for office-based tasks. These projects were summarised at an international conference on telemanipulators for physically disabled people, held in Rocquencourt, France, in 1978 [1, 2]. It is important to note that the first paper of the conference reported a user survey; the most important criterion for success of any rehabilitation equipment is that it should be willingly adopted by users.

An outline of the development of Rehabilitation Robotics in Europe from the Rocquencourt conference to the establishment of the TIDE initiative is presented in the next section of this paper. During that time, the Spartacus work evolved into the French MASTER [3] and Dutch Manus [4] projects, the UK RTX robot [5] was introduced into the market and a number of pioneering University research projects, including some prosthetic limb and gripper projects, developed many aspects of Rehabilitation Robotics theory and practice. Trial installations for office- and factory-based work were studied in The Netherlands, Sweden and the UK.

In general, the information gained in these projects was widely disseminated. The UK Department of Trade and Industry held conferences on Medical and Healthcare Robotics in 1988 and 1989, as part of the brief given to the UK to act as lead country in that field within the International Advanced Robotics Programme. The International Conference on Rehabilitation Robotics (ICORR), which can be regarded as a direct successor to the Rocquencourt conference, began in 1990. A series of workshops on Rehabilitation Robotics was initiated at Cambridge University in 1989 and similar Nordic workshops were held in 1992 and 1993.

TIDE was set up as a pre-competitive technology research and development initiative specifically aimed at stimulating the creation of a single market in rehabilitation technology in Europe. For the purposes of TIDE, rehabilitation technologies are defined as those provided directly to elderly and/or disabled people, to enable them to live more independent lives and become integrated in the social and economic activity of their communities, preferably outside institutional care [6]. It offered the opportunity of collaboration across national boundaries within the fragmented market of rehabilitation equipment and thus opened the way for the many small firms in the field to exploit the benefits of the single European market. It is important to stress that it is a user-led initiative, with user surveys forming an integral part of its projects.

This paper discusses the place of TIDE in the research and development programmes of the European Union and goes on to describe the TIDE projects in, and associated with, the field of Rehabilitation Robotics. The TIDE pilot phase started in 1991 and was succeeded by a bridge phase in 1993. Robotics research and development has been represented in each phase. Among the projects outside TIDE are some supported by other EU funding, including the IMMeDIAte project on integrated wheelchair systems in the SPRINT programme and the PMMA project in the ESPRIT programme.

Prior to the launch of the TIDE pilot phase, a survey was made of the European market in assistive technology. This TIDE market survey provided a snapshot of the needs in rehabilitation research and development by a set of European experts. In the field of control technology, the survey identified a number of potential project areas, including compatibility between input devices and assistive devices in general, robotic aids in vocational environments, the safety of robotic aids, and the design of intelligent wheelchairs including wheelchair-mounted manipulators. While not falling directly under the heading of Rehabilitation Robotics, work on the development of smart wheelchairs is regarded in this paper as an application of mobile robot technology.

Within the TIDE pilot phase, the M3S [7] and RAID [8] projects developed some of the above themes, and the MARCUS [9] project was concerned with the development of a prosthetic hand. A summary of the TIDE pilot phase projects was presented at the TIDE Congress, held in Brussels in April 1993. The TIDE programme is now in its bridge phase, linking the pilot phase to the main phase. The main phase will start in 1995 and will be contained within the 4th EU research framework, under the general heading of Telematics Applications. The term telematics refers to systems and services based upon telecommunications and information technologies.

To provide a more complete picture of the European scene in Rehabilitation Robotics, some of the projects outside the TIDE programme are discussed in the following section. Funding for these projects has been obtained from a variety of sources. Some have been supported by charities in the field of rehabilitation equipment, others have been supported by national research councils.

There are some systems which are already on the market or have an established place in rehabilitation programmes. The eating aid robot, Handy 1 [10], has been successfully developed and marketed as a dedicated application of robot technology. The smart wheelchair programme at the University of Edinburgh CALL Centre [11] has established the feasibility of intelligent wheelchairs as rehabilitation equipment. The Manus wheelchair manipulator and its associated programming environment is being manufactured and marketed by Exact Dynamics bv in The Netherlands.

The final section of this paper discusses the consequences of the TIDE programme in Rehabilitation Robotics, the benefit to the users and the potential for commercial exploitation of the hardware and software arising from the projects. Product exploitation is not the only useful outcome and other projects are leading to the development of standards which will open the possibility of integrating devices from different manufacturers within a single system. As the technology becomes more widely established, increasing effort will be required to establish the associated training and service support networks which are essential to its success.

2 History

A Introduction

European research and development in the field of Rehabilitation Robotics was established in the mid-1970s with the Spartacus and Heidelberg manipulator projects. Since then, activities have been confined to a few well-funded R&D projects, supplemented by several smaller university-based research projects. The availability of a low-cost commercial robot arm with which to conduct research has been significant in this respect. More recently, the EU TIDE programme has provided an opportunity for several strands of research and development to come together in a coherent manner (figure 1). This section outlines some of the earlier projects and their connections with more recent initiatives.

Figure 1. Affiliations between past and current developments in European Rehabilitation Robotics (dates are approximate only)

B Spartacus

Spartacus was a French project involving a feasibility study on the use of a telemanipulator by individuals with high spinal cord injuries [1, 12]. An MA-23 nuclear telemanipulator was used for early experimental work, paying particular attention to control ergonomics. Knowledge gained through the use of this arm led to the implementation of a further system using a custom-designed telethesis (MAT-1). The MAT-1 had six active degrees of freedom powered by motors in the base unit via cable and pulley transmissions. This arrangement led to a compliant robot which would yield to external forces where necessary. One of the most significant conclusions of the Spartacus project was that the effectiveness of a manipulator for rehabilitation would be increased by mounting it on a wheelchair.

C The Heidelberg manipulator

An early example of the workstation-based approach to the implementation of robotic systems was developed at the University of Heidelberg, Germany, in a project sponsored by the German Federal Ministry of Research and Technology [2]. The specification of this workstation was based on the statements of 75 tetraplegic individuals associated with a local rehabilitation centre. The robot arm was under direct control of the operator at all times. At the request of potential users, no pre-programmed motion sequences were provided. A general-purpose pneumatic end effector was used for all tasks with the exception of switch operation and page turning, for which a separately controlled vacuum 'finger' was provided.

D The Dutch Muscular Dystrophy Association manipulator

An experimental wheelchair-mounted manipulator was implemented in 1982 as a private initiative in conjunction with the Muscular Dystrophy Association, The Netherlands [13]. A Cobra RS1 robot arm was mounted on a lap tray and controlled using a keypad. The manipulator was evaluated over a period of three years in a domestic environment. The most frequent tasks performed were eating using a modified spoon, page turning using the rubber tip of the end effector and the manipulation of light switches. This project had a significant influence on future developments in wheelchair-mounted manipulators and on the Manus system in particular.

E Manus

Manus was a collaborative Dutch project started in 1984 and centred at the Institute for Rehabilitation Research (IRV) [4]. Manus built on the conclusions of the Spartacus project by specifying and developing a wheelchair-mounted manipulator. The original Manus manipulator consisted of a 5 degree of freedom arm mounted on a rotating and telescoping base unit which could be attached to a variety of electric wheelchairs. Current versions have a reach of approximately 850 mm and can lift 1.5 kg at full extent. The use of slip couplings and a firmware watchdog increase the safety of this manipulator. Manus has been extensively evaluated within rehabilitation centres over several years and is now being manufactured and marketed by Exact Dynamics bv, The Netherlands. The Manus arm has been used within the TIDE M3S project to demonstrate the functionality of a proposed wheelchair-based communication bus standard.

F The RTX robot arm

The RTX robot arm has been influential in much of the research and development in Rehabilitation Robotics around the world. It was designed by Universal Machine Intelligence Ltd, UK for light industrial, rehabilitation and laboratory applications and first came to market in 1986 [5]. A survey published in 1989 showed that 38% of all workstation-based Rehabilitation Robotics projects world-wide were using the RTX [14]. The arm is of SCARA geometry and has dimensions approximating those of an adult. Following the demise of UMI, rights to the RT series robots were acquired by Oxford Intelligent Machines Ltd, UK and the arm continues to hold a unique market position in terms of price and performance. Upgraded versions of the robot have become available recently featuring a Transputer-based motor controller. These upgrades should extend the scope of the robot into more demanding application areas. A derivative of the RTX robot arm has been used within the TIDE RAID and EPI-RAID projects as a manipulative tool for disabled office workers.

G MASTER

Development of the MASTER workstation was started in 1985 by the French Atomic Energy Commission (CEA) and draws on experience gained within the earlier Spartacus project. The system uses an RTX robot arm with modified control electronics and interchangeable end effectors [3]. The arm is programmed using a teach pendant and direct manipulation modes are also available. Environmental control facilities form an integral part of the MASTER controller. The prototype systems were commanded via a scrolling menu system displayed on an LCD panel in conjunction with a speech recognition system. They have been evaluated at a number of French rehabilitation centres. More recently, the MASTER II system has been developed, employing a Windows-based human-computer interface running on a PC-compatible computer. Control technologies from the MASTER workstation have been further utilised within the TIDE RAID and EPI-RAID projects.

H CURL

The Cambridge University Robot Language (CURL) is a task-level programming environment designed specifically for rehabilitation applications [15]. It was conceived in 1987 following a number of trials using the RTX robot arm to perform tasks programmed with a conventional compiled language. The perceived inadequacies of this approach led to the specification of CURL. The programming language is interactive and employs a natural language syntax to encourage the development or modification of tasks by non-technical personnel. The environment also incorporates a world model to facilitate programming at the task level. The combination of these two features allows CURL commands of the form move the book to the shelf. More recently, CURL has been ported to the Microsoft Windows environment and restructured to provide robot-independent operation. The interfacing of CURL to a specific robot is achieved through a CURL device driver coded to an open specification. CURL has recently attained commercial status and has been specified as the robot command language for the next generation of RAID workstations to be developed within the TIDE EPI-RAID project.

I HADAR workplace adaptations

The work of HADAR, Sweden serves as a good example of the many robotic workplace adaptations which have been undertaken in recent years. These adaptations are mainly confined to the Nordic countries, where substantial finance is available from national government for the vocational rehabilitation of individuals sustaining severe injuries. One of the first permanent workplace adaptations to feature robot technology was implemented in 1990 by HADAR for invoice processing work. This system consisted of an RT-series robot arm with a modified end effector employing suction cups to hold paper documents. More recently, an installation at Samhall-HADAR has been used for the transfer of old manuscripts to a computer system [16]. These installations both employ CURL to command the robot in conjunction with a simplified user interface.

3 Introduction to TIDE

A The EU role in European R&D co-ordination

The European Union organises, according to its Treaties, all its research and technology development (RTD) work in a 5-year plan, which is called the Framework Programme. For the next few years this will be the Fourth Framework Programme (IV FP). The objectives are generally to help the European economy to be competitive and to secure the best prospects for its future growth for the benefit of its citizens. Keywords are therefore 'industrial competitiveness' and 'sustainable economic growth' which lead to a high level of employment. A key phrase is also 'quality of life'. Specific emphasis is given in the IV FP to the modernisation of the European economy through information and communication technology, aiming at the so-called information society, to ensure access by small and medium enterprises to advanced technologies, and also to the co-ordination and complementing, with respect to the subsidiarity principle, of national efforts and European Union-wide efforts [17].

The IV FP is organised into four activities: firstly research, technological development and demonstration projects, secondly co-operation with non-EU countries and international organisations, thirdly the dissemination of research results and preparation for exploitation, and finally the stimulation of training and mobility of researchers. The budget of the first activity is 9.4 billion ECU of European Union funding. Within the first activity, a major domain is information and communication technologies, in which three programmes reside, namely the Telematics Applications programme with a budget of 843 million ECU (MECU), the Advanced Communications Technologies programme with a budget of 630 MECU, and the IT programme with a budget of 1932 MECU [18].

B Telematics applications

Telematics Applications is an RTD programme of applications of information and telecommunication technologies that takes into account the new opportunities offered by areas such as multimedia, HDTV, interactive TV and virtual reality. It contains the sub-programme TIDE with a budget of 65 MECU.

C TIDE's objective and approach

TIDE's objective is to promote assistive technology RTD to meet a social and an industrial goal, namely to improve the quality of life for disabled and elderly persons, and to improve the European industry and market in products and services that meet the needs of disabled and elderly persons [19].

TIDE proposes a two-way approach to meet this objective. The first one is to design general consumer products and services in such a way that they become accessible or usable by disabled and elderly persons, which is called the 'design for all' approach. The second approach is to develop special devices and services for disabled and elderly persons for the compensation of impaired functions and for special interfaces to commonly used equipment. In short, assistive technology or rehabilitation technology [20, 21].

D The history of TIDE

Assistive technology is a field in its own right, that brings together a multitude of disciplines including engineering, informatics, materials technology, psychology, education, sociology, and medical science. Examples of advanced assistive technology products and services, as being developed within TIDE, are intelligent living environments for elderly people, neural net-based hearing aids, fuzzy-logic controlled self-navigating wheelchairs, robotic aids, speech recognition and speech output devices, alert/alarm systems, access to graphical computer software for blind persons, and virtual environments for helping people to restore their mobility after an accident or for training children with communication disorders.

The TIDE initiative started with a pilot phase in 1991. At that time, 21 RTD projects were launched in a wide variety of domains, with EU support of 18 MECU. The pilot phase was extended in 1992 by 10 projects and a large study into standardisation, legislation and economic aspects, service delivery, industrial structure, training and new technologies [22]. In 1993 the TIDE bridge phase was launched, which led to a further 55 projects and horizontal activities, running from 1994 to 1996, with a total EU funding of 42 MECU [19]. TIDE will be continued into a main phase under the umbrella of the Telematics Applications programme, which is expected to lead to new projects from 1995 onwards. The new workplan is expected to contain, next to horizontal activities and accompanying measures, the following areas of RTD work:

In the area of Rehabilitation Robotics and advanced wheelchairs, future RTD work has been proposed in areas such as 'go anywhere' wheelchairs, in-house transfer systems, intelligent manipulator systems (including assessment of current systems), navigation and docking systems and domestic intelligent assistants.

TIDE projects are run by consortia of industrial and research organisations, service deliverers and users (represented though their professional organisations and rehabilitation centres). Generally, each consortium must contain at least two industrial partners from different European Union or EFTA countries. The contribution from the European Union to the funding is up to a maximum of 50% of the full cost for the commercial partners and can be up to 100% of the additional cost for participation of other EU partners.

Project proposals are judged on a range of criteria, amongst which user involvement and commercial relevance of the proposal play a key role, next to the technical and managerial soundness.

E The significance of TIDE in European Rehabilitation Robotics

TIDE has attracted a lot of attention in Europe, and it is probably fair to say that it has become a focus for European assistive technology. The initiative has an extremely good participation of small and medium enterprises and organisations (about 74% of all proposers in the bridge phase). There is also substantial participation by non-profit organisations and public service providers. Recently, large industry is starting to show more interest for this field too, motivated by the prospects of the market for the older population.

However, the goals of TIDE are not yet fully realised. In particular, for many segments of the assistive technology market there does not yet exist a 'single European market'. In general terms, the product offer is rather limited, confined to national boundaries, and the price/performance ratio is often rather unfavourable. There are many reasons for this, and extensive studies examine this (dis-)functioning of the market in more detail [23, 24]. Fragmentation by country and by disability are not the least amongst these reasons. There exist many barriers to achieving the economies of scale of the larger market. These include language and cultural differences, country-specific certification, lack of transparency of distribution and service delivery systems, and incomplete information flow.

These non-technical barriers also play a role in the Rehabilitation Robotics market, which additionally has to deal with the market initiation problems of any truly innovative domain. Specifically, it is observed that in Europe:

  • Rehabilitation Robotics is technically advanced, with a lot of technical development and potential commercial benefits, but so far little market impact
  • Rehabilitation Robotics in Europe is addressed by a limited number of groups
  • The danger exists of getting stuck to the image of the 'ever to remain' future technology
  • There is little co-operation among manufacturers and no trade association exists; there is no forum where manufacturers, researchers, rehabilitation centres and user representatives meet
  • The current state of the market is very much in an initial phase:
  • The field has a relatively low profile with respect to support from EU and national funds In short, the field of Rehabilitation Robotics in Europe, although a quite active field with a good technical and commercial potential, suffers from fragmentation and lack of scale at the national level. There is therefore good reason to support the field through European initiatives such as TIDE.

    Collaboration in TIDE projects should lead to new synergies, for example bringing technologies together that have been developed in several countries and are jointly needed to provide new solutions. TIDE projects also aim at improved effectiveness, in the sense that solutions have a greater chance to be ultimately exploited for the benefit of users. In particular, TIDE enables industry and users (in this case rehabilitation centres) to come together at a sufficiently large scale in order to validate solutions. TIDE also aims to improve efficiency in the field, for example by promoting solutions that can address the larger European market with its economies of scale.

    4 TIDE projects

    A Introduction

    The TIDE initiative addresses many disabilities, and correspondingly a broad range of assistive technology tools are the subject of the TIDE projects. Robotics and advanced wheelchairs (in the sense of 'intelligent' platforms) are being addressed in 4 of the 21 TIDE pilot phase projects, and in 5 of the 55 TIDE bridge phase projects. These projects fall roughly into two categories: those that focus at the functional component level of a general robotic aid architecture (eg a manipulator module), and those that provide an integrated robotic aid solution for one or more selected applications, eg a workstation for a motor-disabled DTP operator (table 1).

    Functional components
    Integrated solutions
    Manipulators
    Navigation
    Technology for integration
    Robotic workstations
    Integrated platforms
    Pilot phase
    MARCUS
    -
    M3S
    RAID
    MECCS
    Bridge phase
    -
    SENARIO
    FOCUS
    EPI-RAID
    OMNI and MOVAID

    Table 1. Classification of TIDE projects in Rehabilitation Robotics

    B MARCUS

    The original target user group of MARCUS are persons who need a hand prosthesis. Within the TIDE pilot phase, the MARCUS project developed a novel artificial hand that offers the user a more natural control of grip and grasp operations. In particular, the hand has been equipped with force and slip sensors and incorporates a feedback control mechanism that takes over much of the user control required for other prosthetic hands. In addition to enhanced functionality, major requirements of prosthetics users are comfort and appearance. These have been addressed by an innovative anthropomorphic mechanical design and a new type of glove. The artificial hand as a prosthesis is a self-contained system that is interfaced for user control through myo-electric electrodes.

    The hand also has applications as a component in the field of Rehabilitation Robotics. In particular, it is now applied as the manipulator part of the MOVAID integrated system described below.

    C SENARIO

    The SENARIO project is developing a sensor aided intelligent navigation system that can be mounted on a range of powered wheelchairs. As such, it is providing the navigation component for a fully integrated mobile and possibly robotic system that potentially has applications outside the field of rehabilitation. The system will consist of hardware and software sub-systems (some of which may become independent products) for sensing, risk avoidance (including sensor fusion software), positioning, user control, power control, as well as interfaces and interaction modules between the sub-systems. The M3S bus is an option for interfacing between the control panel, central risk avoidance and power sub-systems. The main sensor is ultra-sonic, but other types of sensors, such as odometers and inclinometers, will also be integrated. The positioning system will use a range finder with active and/or passive beacons.

    The system can be either in an off-line teaching mode (to teach pre-defined routes based on a topographical representation of the vehicle's workspace), or in a run mode. In the latter case, either predefined routes are followed by autonomous navigation or, by semi-autonomous navigation, a free route is followed under user control with assistance from the risk avoidance sub-system. Clearly safety is one of the most important user requirements and a basic target for the project.

    D M3S

    M3S is an interface and bus specification for input devices and end effectors in rehabilitation equipment. Wheelchair motor controllers, manipulators, robot workstations, environmental controllers and voice synthesisers are examples of end effectors. The project is therefore not specifically robotic, but has produced an enabling technology through which devices, including robots and manipulators, from different manufacturers may be connected together in a single system while maintaining safe operation and full compatibility.

    The acronym M3S stands for Multiple Master Multiple Slave. The specification describes:

    The bus uses the CAN (Controller Area Network) communication protocols, plus circuits and wires for extra safety procedures. Information is exchanged between devices in digital form using message packets on the two-wire CAN bus. The specification allows a system to be optimally configured to the user's needs. The intelligence in the system exists in the controllers of the individual end effectors and also in the control and configuration module (CCM) which forms part of every implementation [25].

    For M3S trials, a version of the Manus manipulator has been integrated into a wheelchair system operating from a single input device. Preliminary user trials have been carried out on the selection and operation methods in the specification.

    E FOCUS

    The FOCUS project in the TIDE bridge phase is a successor to the M3S pilot phase project. As in the M3S project, a large consortium of industry, research, rehabilitation centres and user associations has been set up in order to gain the widest possible support from the very start.

    The viability and ease of use of the M3S architecture will be improved in FOCUS through activities to promote de jure as well as de facto standardisation of M3S by providing chip designs for the key elements and by providing design rules for safety. The M3S specification will be expanded, and specially designed chips for M3S interfaces will be made available at low cost. Ultimately the goal is to enable the provision of integrated wheelchair - robotic aid - environmental control systems that can be optimally configured for the individual user at relatively low cost through a modular approach and the use of the M3S standard.

    The functionality of demonstration platforms constructed in the project will be extended compared to the previous project in terms of an improved user interface, navigation assistance and an infra-red link and bridge to home buses. These demonstrator platforms are integrated systems consisting of a powered wheelchair equipped with the M3S bus, navigational control with a safety zone, a wheelchair-mounted robotic arm, a choice of input devices (sip and puff sensor, headrest control sensor, finger sensor, switches, joystick, keyboard and scanners) and user interface hardware and software for integrated control. M3S is also used in the TIDE projects MOVAID and LAMP, and in the SPRINT project IMMeDIAte.

    F RAID

    The RAID workstation was developed to provide manipulative assistance to severely physically disabled individuals within office environments [8]. The system features an upgraded RTX robot arm mounted on a linear track. The arm accesses books, paper documents, diskettes and CD-ROMs placed on an integral shelving system. The operator commands the robot from the same computer on which office work is performed. Books, paper documents and printer output may be brought from the shelves to a reader board situated on the operator's desk. An additional page-turning gripper allows the written information to be browsed. The original RAID design was targeted at CAD operators due to the high quantity and physical size of the reference materials which must be accessed. Trials of the prototype system revealed a requirement for a smaller workstation and for the application of the RAID concept in more common vocational activities.

    G EPI-RAID

    The EPI-RAID project builds on the concepts of the prototype RAID workstation. The design has been modified to provide greater reliability and a degree of system modularity. Further trials are planned prior to the specification and development of a second generation workstation. The RAID 2 workstation will feature the latest RT200 robot arm and will be programmed using CURL. It is intended that intelligent grasping and self-configuration routines will be incorporated to reduce the complexity of calibration and task programming activities. RAID 2 will also be suitable for use in domestic environments where the nature and diversity of tasks cannot be exhaustively analysed at the specification stage.

    H MECCS

    The MECCS pilot phase project is included here in the category of integrated platforms. Its approach is to integrate environmental control on the wheelchair platform, interconnecting to a home bus. Although it did not include a robotics component it is relevant in this context because of the potential further development of the system to provide a user interface to the M3S bus and thereby control robotic aids too. As such, MECCS provides elements for composing a completely integrated mobile platform with robotic devices. The system consists of a laptop computer and radio communications module on the mobile base, as well as a fixed radio module interfaced to a target home control bus. A series of alternative input devices can be connected to the laptop. The project has resulted in a more profound understanding of the requirements of mobility impaired customers with respect to environmental control.

    The approach of MECCS is to provide a controller terminal, mounted on the wheelchair, capable of allowing the use of control devices in the immediate environment via a radio gateway link to a Home Bus control system. The system is designed to permit the easy integration of augmentative communication systems for mobility control and to provide a further gateway to control the wider environment outside the house.

    I OMNI

    The OMNI bridge phase project provides a major step forward in natural wheelchair control by persons with a severe physical or multiple disability by providing omni-directional mobility of the wheelchair. This allows the chair to move in any direction and the linear motion to be combined with a rotation around any given point, including the centre of the wheelchair.

    In addition to the omni-directionality, which facilitates intuitive user control and increases mobility, ultra-sonic and infrared sensors for environment analysis with an obstacle-avoidance system are provided. This enables high-level control of the wheelchair navigation, reducing the complexity of the control task for the user while guaranteeing safety. The system is equipped with a modular human-machine interface which supports a variety of input devices and levels of user abilities (from accurate hand control of mouse or joystick to switch controlled auto-scanning mode). Environmental control and an elevating seat complete the system. The user interface is uniform for the different applications of direct wheelchair control, control via the navigation module and environmental control.

    The system provides a versatile and highly mobile platform which is suitable for further extension with other end effectors such as robotic devices. Whereas vocational rehabilitation is addressed at this stage, future application in a domestic environment can be readily conceived.

    J MOVAID

    The objective of the MOVAID project is to give disabled and elderly users more comfortable access to and control of non-specialised consumer products such as food preparation or house cleaning equipment. For severely disabled or bed-ridden users, the solution is a modular mobile robotic assistance system that interacts with activity workstations. For other users, this consists of a range of user interfaces for appliances. The latter are possibly M3S based, and as such can also be controlled from other M3S-based systems such as a robotic base. Explicit in the MOVAID approach is that the user continues to play an active role in the control of the environment, through mediation of the robotic system. The modular robotic system features a mobile base equipped with an innovative 8 degree of freedom robot arm with low-level controller and the MARCUS gripper, sensory systems for navigation and obstacle avoidance, a docking capability to activity workstations, off-line local control sequences and remote control, optionally with video inspection. The activity workstations can be adapted by users who already own a robot arm and facilitate the high-level control of the arm. A major component of the project is the design of the user interfaces for the workstations and for the control of the robotic system.

    5 Non-TIDE projects

    Outside the TIDE programme, there are a number of European research and development projects in the field of Rehabilitation Robotics. This section describes the most significant of these projects. Funding has been forthcoming from a variety of sources including:

    A Handy 1

    The Handy 1 is a robotic aid conceived and developed at Keele University, UK. The first prototype was developed in 1988 specifically to assist a child with Cerebral Palsy in eating independently [10]. The aid features a 5 degree of freedom robotic arm and a food tray mounted on a wheeled base unit. The arm is controlled by a single switch input device in conjunction with a number of LEDs which illuminate in sequence. Over 60 units have now been placed with individuals of varying ages and disabilities for evaluation. The Handy 1 has been shown to improve the eating skills of regular users over time due to the consistency with which food is presented. More recently, the aid has been considered for other activities including drinking, shaving and teeth cleaning. The device has won several awards for innovation and design. The manufacture, marketing and service of Handy 1 units is now undertaken by Rehab Robotics Ltd, UK.

    B IRVIS

    The Interactive Robotic Visual Inspection System (IRVIS) is the product of an investigation started in 1989 at the University of Cambridge, UK into the application of interactive robotics as an assistive technology in manufacturing engineering [26]. The project was jointly funded by the ACME directorate of the UK Engineering and Physical Sciences Research Council and the Papworth Group, UK. Criteria by which the potential of assistive robotic devices in specific vocational task areas could be assessed were identified. The visual inspection of electronic circuits was then selected as a suitable task on which to base further study. The IRVIS was developed in consultation with personnel at a hybrid microcircuit manufacturing plant. It features a 5 degree of freedom mechatronic system supporting a high magnification CCD video camera and a circuit tray. The operator adjusts the relative positions of the camera and circuits from a PC compatible computer using the Cambridge University Robot Language (CURL). A six month site trial of the IRVIS with a disabled operator has been undertaken. The trial has revealed limitations in the method of user interaction and reliability issues which will be addressed in a continuation project.

    C IRQAT

    In parallel with the IRVIS project, the Papworth Group has been developing an Interactive Robot Quantitative Assessment Test (IRQAT) [27]. This test measures a person's cognitive ability to use an interactive robotic workstation using a 'peg in hole' task. This task was selected because it is easy to specify and conceptualise. It has also been adopted in previous related research. The quantitative assessment of an individual's ability and aptitude to use such equipment is seen as important in the creation of employment opportunities. The test employs direct control and task-level control techniques such that the most appropriate form of interaction for an individual may be determined.

    D The InventAid arm

    The InventAid Arm is a wheelchair-mounted manipulator. It uses patented pneumatic actuators based on the Flexator air muscle. The manipulator can lift 3 kg to a height of 1.2 m and is sufficiently sensitive to grasp an egg. It can be retrofitted to most electric wheelchairs and intentionally uses no electronic components. The InventAid arm may therefore be serviced within most workshops. A small number of these manipulators have been manufactured by the Papworth Group, UK and are being evaluated [28].

    E The Wessex arm

    The Wessex arm is the product of on-going research and development in Rehabilitation Robotics which started in 1986 at the Bath Institute of Medical Engineering (BIME), UK. It is a trolley-mounted arm of SCARA geometry designed to be moved around domestic environments with its operator [29]. The arm may be controlled directly as a telemanipulator or can be taught motion sequences for subsequent replay. Work on the arm started with a feasibility study including a user survey, interface trials and prototype robot trials. This study led to the design of a workstation-based robotic aid and trials at a local hospital. The aid was subsequently refined to produce the current trolley-mounted version.

    F Tou

    Tou is a robot arm under development in a recent project at the Polytechnic University of Catalunya, Spain [30]. The robot is designed to be intrinsically safe due to a structure which is both soft and highly compliant. Tou is constructed from cylindrical segments of foam rubber. Each segment may be deformed in 2 degrees of freedom using cables pulled by motors in a base unit. The arm is intended to complement rather than replace the assistance provided by carers. Operators may use Tou to carry out tasks such as page turning and scratching where great strength and precision are not required. The arm is normally controlled directly, however, a number of basic motion sequences may be programmed into the controller. Tou has been evaluated by a number of tetraplegic individuals at a rehabilitation centre and has been welcomed as an experimental tool. Current work includes the integration of a vision system to improve performance and functionality.

    G URMAD

    URMAD is an Italian acronym which means 'Mobile Robotic Unit for the Assistance of the Disabled'. This robot is being developed by a consortium of universities and industrial companies and is funded through the National Research Council of Italy [31]. URMAD consists of an 8 degree of freedom arm mounted on a mobile base and controlled from a static workstation via a radio link. The arm incorporates a force/torque sensor and is highly compliant to minimise the potential for injury. URMAD is primarily a research platform and incorporates many 'state of the art' technologies including vision, ultrasonics, and speech recognition. Short-term objectives include navigation in semi-structured environments, visual recognition of known objects and object grasping. Techniques developed within the URMAD project are already being utilised elsewhere within the TIDE MOVAID project.

    H WALKY

    WALKY is a mobile robot system designed to assist physically disabled people to work in a chemical laboratory [32]. This vocational field has been identified as a potential alternative to office activities for the application of robotic assistive technologies. The system is being developed at Lund University, Sweden. WALKY consists of a Scorbot ER VII robot arm mounted on a TRC Labmate mobile base and a number of sensing sub-systems employing infra-red and ultrasonic transducers. The combination of these technologies allows an operator to retrieve an object from a remote workplace. The robot will avoid obstacles where possible and uses a combination of pre-programmed procedures and direct control techniques to grasp objects. The high-level controller uses a CAD drawing of the floor plan both as a form of user interface and also to assist in navigation.

    I The Middlesex arm

    Researchers at Middlesex University, UK have made a thorough investigation into the design of a wheelchair-mounted robotic manipulator. This work commenced in 1988 with a survey of 50 electric wheelchair users across the UK which established a need for such a device for the purpose of manipulating objects and reaching to the floor [33]. The use of muscle-type actuators for robotic aids has also been analysed. A novel kinematic arrangement for a wheelchair-mounted robotic aid has been designed. This new arrangement includes aspects of SCARA geometry with an additional axis extending the working envelope to floor level. Testing of a prototype arm will commence in the near future.

    J The Smart Wheelchair

    The Smart Wheelchair, developed at the University of Edinburgh, UK, was conceived as a tool for the investigation of augmentative mobility in the education and therapy of multiply disabled children [11]. Sensors for proximity, line following, odometry and collision detection facilitate adaptation to the requirements of many users. The system continues to evolve as experiments with the system reveal new possibilities. A key to the success of the Smart Wheelchair lies in its highly modular software design using an object-oriented Forth environment and a behaviour-based approach to sensor integration.

    K IMMeDIAte

    The IMMeDIAte project involves a total of 17 collaborating rehabilitation centres, equipment manufacturers, component suppliers and research organisations in the realisation of 5 different integrated rehabilitation systems. Each system includes a wheelchair, an environmental control system, and a navigation sub-system. The sub-systems are interconnected using the M3S communication bus. IMMeDIAte is funded within the EU SPRINT programme. Systems incorporating either the Manus manipulator or the MASTER workstation feature in the project. The validity of such highly integrated systems will be evaluated in a subsequent trials phase. Finally, a demonstration phase will involve presentation of the technology across Europe.

    L PMMA

    Within the EU ESPRIT programme, a working group has been established to specify a system of integrated assistive devices for future applications in rehabilitation. The proposed system has been described as providing Personal Mobile Manipulation Assistance (PMMA). State of the art technologies such as behaviour-based robotics and semi-autonomous navigation are candidates for inclusion in the specification.

    6 Summary and conclusions

    A TIDE in practice

    A considerable number of Rehabilitation Robotics and related projects are going on in Europe, partially with support of European Union funds. As far as technological development is concerned, results look promising. There is a rich flow of technical information from related fields into the field of Rehabilitation Robotics and further rapid technical progress seems quite feasible. Insight into user requirements and into the limitations and capabilities of robotic systems to provide real solutions to users is steadily increasing, although the exchange of user experiences for example between rehabilitation centres could be further stimulated.

    The conclusion, as far as commercial exploitation is concerned, seems to be that many of the systems are delivered as prototypes only and do not make it to the market. Some reasons for this under-exploitation have been indicated above. They have to do with the structure of the European Rehabilitation Robotics market, in terms of national fragmentation, lack of distribution and service channels, financing, information flow, and co-operation and competition amongst actors. Specifically, we could argue that existing Rehabilitation Robotics manufacturers would benefit from more market competition in that a more serious investment in both marketing and technology would be required. An extended range of solutions would lower the barriers to product acceptance, and the increased exposure of solutions to the public and purchasers would stimulate diffusion. European programmes can also play a role here, especially where they offer possibilities to support dissemination of information, technology transfer and access to venture financing.

    B Future development

    There are two general factors which are likely to influence the future development and exploitation of Rehabilitation Robotics technology [34]. These arguments are not unique to development within Europe, but are discussed here in the European context: The availability of funds to provide disabled people with workplace support will have a considerable influence on the future prospects of vocational robotic workstations such as RAID and those developed by HADAR in Sweden. A growing awareness of the potential capabilities of many people with physical disabilities may lead to improvements in the funding procedures for assistive equipment and open increased opportunity of employment. These changes will carry with them the need for employers to invest in facilities for training and employing disabled workers. A related factor is increasing awareness of the need, which may be influenced by legislation, to provide wider access to computer-based tasks within an office environment by the widespread inclusion of hooks allowing special-purpose access methods to be used in generic programming systems.

    Despite the employment opportunities opened in this way, the market for this form of workstation is not large. The robot assistant will therefore remain an expensive solution to the problems of picking and placing books, paper and other objects in the workplace. The scope for other forms of robotic assistance is equally limited. Nevertheless, the HADAR experience shows that disabled people, with individually adapted work-aid devices and support, are a resource and an asset in themselves for the labour market [35].

    Experiences with Manus, the main example now available of a robotic system for use in an unstructured domestic environment, show that it provides a significant improvement in a user's ability to perform everyday functions more independently [36]. While it seems unlikely that there will be mass-production of domestic or light industrial interactive robotic assistive devices for general-purpose use by disabled people, the increasing awareness of the possibilities of both manipulators and autonomous vehicles is likely to open new opportunities for domestic applications of robotics.

    The technical changes, linked to smart house technology, which are slowly beginning to make their presence felt in the domestic appliance market, will have considerable benefit in the field of assistive technology in the activities of daily life [37]. Also, and with wider benefit, the smart house concepts should mean that more mass-production standard domestic appliances will either be accessible to people with disabilities or, as in computer technology, contain hooks making them accessible via additional interface devices. The evolving M3S specification will be of considerable importance in this field if it can provide a low-cost and standard means of giving wheelchair users a link into smart house systems.

    Bearing in mind the limitations discussed earlier in this section, it seems likely that progress in robotic assistive technology in the immediate future will continue to rest with a modest number of small-scale developments. However, much has been achieved to date and current developments indicate promising possibilities for the future.

    Appendix

    The following acronyms have been used in this paper, note that certain European acronyms have no direct English translation:

    Acronym Interpretation
    CAN Controller area network
    CURL Cambridge University Robot Language
    ECU European currency unit
    EFTA European free trade association
    EPI-RAID Evaluation of prototype and improvements to RAID workstation
    ESPRIT European strategic programme for research and development in information technology
    FOCUS Focus on the central position of users in integrated systems
    ICORR International conference on Rehabilitation Robotics
    IMMeDIAte Integrated system for mobility and manipulation for disabled people
    IRQAT Interactive robotic quantitative assessment test
    IRVIS Interactive robotic visual inspection system
    M3S Multiple master multiple slave
    MARCUS Manipulative automatic reaction control and user supervision
    MASTER Manipulator autonomous at service of tetraplegics for environment and rehabilitation
    MECCS Modular environmental control and communications system
    MOVAID Mobility and activity assistance systems for the disabled
    OMNI Office wheelchair with high manoeuvrability and navigational intelligence for people with severe handicap
    PMMA Personal mobile manipulation assistance
    RAID Robot for assisting the integration of disabled people
    RTD Research and technology development
    SCARA Selective compliant articulated robot for assembly
    SENARIO Sensor aided intelligent wheelchair navigation
    SPRINT Innovation and technology transfer
    TIDE Telematics for improving the quality of life of disabled and elderly people
    URMAD Mobile robotic unit for the assistance of the disabled

    References

    1. Guittet J, Kwee HH, Quetin N, Yclon J (1978) The Spartacus telethesis: Manipulator control and experimentation. IRIA Proceedings of the International Conference on Telemanipulators for the Physically Handicapped. 79-100
    2. Roesler H, Küppers HJ, Schmalenbach E (1978) The medical manipulator and its adapted environment: A system for the rehabilitation of severely handicapped. IRIA Proceedings of the International Conference on Telemanipulators for the Physically Handicapped. 73-77
    3. Cammoun R, Détriché J-M, Lauture F, Lesigne B (1993) Clinical evaluation of the 'MASTER' robot system and development of a new version. Robotica. 11. 535-539
    4. Kwee HH, Duimel JJ, Smits JJ, Tuinhof de Moed AA, van Woerden JA (1989) The Manus wheelchair-borne manipulator: System review and first results. IARP Proceedings of the 2nd Workshop on Medical and Healthcare Robotics. 385-395
    5. Walker N (1986) RTX robot arm. Personal Computer World. Dec 1986, 176-181
    6. TIDE pilot action workplan. Commission of the European Communities, Directorate-General Telecommunications, Information Industries and Innovation.
    7. van Woerden JA (1993) A safe and easy to use integrated control and communication method, M3S. Rehabilitation technology - Strategies for the European Union. IOS Press. 75-80
    8. Dallaway JL, Jackson RD (1992) RAID - A vocational robotic workstation. ICORR 92 - Conference papers
    9. Kyberd P, Tregidgo R, Sachetti R, Schmidl H, Snaith M, Holland O, Marchese S, Bergamasco M, Bagwell P, Chappell P (1993) Rehabilitation technology - Strategies for the European Union. IOS Press. 98-102
    10. Topping M (1994) 'Handy 1' a robotic aid to independence for severely disabled people. Proceedings of the 3rd Cambridge Workshop on Rehabilitation Robotics. 13-16
    11. Craig I, Nisbet P (1993) The Smart Wheelchair: An augmentative mobility 'toolkit'. ECART 2 - Proceedings. 13.1
    12. Kwee HH (1986) Spartacus and Manus: Telethesis developments in France and The Netherlands. Interactive robotic aids - One option for independent living. World Rehabilitation Fund. 7-17
    13. Zeelenberg AP (1986) Domestic use of a training robot-manipulator by children with muscular dystrophy. Interactive robotic aids - One option for independent living. World Rehabilitation Fund. 29-33
    14. Prior SD (1989) A review of world Rehabilitation Robotics research. Internal Report - Middlesex Polytechnic Mechanical Engineering Department.
    15. Dallaway JL, Mahoney RM, Jackson RD, Gosine RG (1993) An interactive robot control environment for rehabilitation applications. Robotica. 11. 541-551
    16. Holmberg L (1993) The installation of a robotized workstation at Samhall-HADAR. ECART 2 - Proceedings. 29.2
    17. Information provided by Commissioner M Bangemann on 8 April 1994, CORDIS FOCUS newsletter, Issue 14, 22 April 1994, Commission of the European Communities DGXIII-D-2
    18. Decision of the European Parliament and the Council. 26 April 1994
    19. Decision of the Council on TIDE. 1993-1994
    20. Timmers P (1994) USER Newsletter. Issue 1. April 1994
    21. Ballabio E, Timmers P (1994) Proceedings of the Italian Conference on Virtual Reality. Bologna, May 1994
    22. TIDE pilot phase synopses, EUR 15023 EN (1993) Office for Official Publications of the European Communities. Luxembourg, March 1993
    23. CORE TIDE pilot phase project, final report (1994) IRV. The Netherlands, April 1994
    24. HEART TIDE pilot phase study - Census of legislation and assessment of impact of legislation on rehabilitation technology (1994) TIDE Office, DGXIII-C.5, European Commission. Belgium, April 1994
    25. van Woerden JA, Nelisse MW, Perricos C, Jackson RD, Davies B, Hibberd RD, Banerjee D (1994) M3S - a standard communication architecture for rehabilitation applications. IEE Computing and Control Engineering Journal. 5. 213-218
    26. Dallaway JL, Mahoney RM, Jackson RD (1994) The application of Rehabilitation Robotics within manufacturing industry. ICORR 94 - Proceedings. 145-149
    27. Dargie G (1994) Handbook for the Interactive Robot Quantitative Assessment Test (IRQAT). Internal report - Cambridge University Engineering Department. CUED/F-INFENG/TR162
    28. Mattie J, Hannah R (1994) Evaluation of the Inventaid manipulator arm for youth with Muscular Dystrophy. ICORR 94 - Proceedings. 13-17
    29. Hillman M, Gammie A, Orpwood R (1993) Design of a trolley mounted robotic aid. ECART 2 - Proceedings. P.I
    30. Casals A, Villà R, Casals D (1993) A soft assistant arm for tetraplegics. Rehabilitation Technology - Strategies for the European Union. IOS Press. 103-107
    31. Guglielmelli E (1994) Mobile robots in residential care. ROBOmed 94 - Proceedings. 89-103
    32. Neveryd H, Bolmsjö G (1994) WALKY, a mobile robot system for the disabled. ICORR 94 - Proceedings. 137-141
    33. Prior SD, Warner P (1994) Investigations into the design of a wheelchair-mounted rehabilitation robotic manipulator. Proceedings of the 3rd Cambridge Workshop on Rehabilitation Robotics. 32-48
    34. Jackson RD (1993) Robotics and its role in helping disabled people. IEE Engineering Science and Education Journal. 2. 267-272
    35. Trulsson B-G, Persson E-M (1993) Coordinated work-oriented rehabilitation and workplace adaptation at HADAR. ECART 2 - Proceedings. 25.1
    36. Øderud T, Bastiansen JE, Tyvand S (1993) Experiences with the Manus wheelchair mounted manipulator. ECART 2 - Proceedings. 29.1
    37. Ferreira JM, Cooper M, Keating D (1994) Home systems technology for elderly and disabled people: Present status, R&D methodology and future directions. RESNA 94 - Proceedings. 45-47

    Biographies

    John L Dallaway

    John Dallaway was educated at the University of London, gaining a BSc in Physics with Microcomputer Electronics and a PhD in Rehabilitation Robotics. His postgraduate studies were centred on the integration of intelligent task-level and direct robot control strategies under a common user interface. Dr Dallaway joined the Rehabilitation Engineering Research team at the University of Cambridge in 1992 and has been involved in several projects concerning the high-level command of assistive mechatronic systems. He is a co-author of the Cambridge University Robot Language (CURL).

    Robin D Jackson

    Robin Jackson is a Lecturer in the Cambridge University Engineering Department and a Fellow of Selwyn College. His current research activities are in the general area of Rehabilitation Robotics and include control methods and applications for physically handicapped persons. He is concerned with the use of robotic devices within workstations in educational and vocational applications. A related research area is the control of electric machine drives. He is a Chartered Engineer, a member of the IEE and a member of the Biological Engineering Society.

    Paul HA Timmers

    Dr Paul Timmers is currently working in the European Commission, Directorate-General XIII, as Assistant to the Director of the Telematics Applications Programme. He has been involved in the TIDE initiative from the launch of the pilot phase onwards, with a specific interest in Rehabilitation Robotics. Before joining the Commission, he held various positions in development, product management and marketing of computer systems. He did his PhD research in the domain of Elementary Particle Physics.

    This paper may be referenced as follows:

    Dallaway JL, Jackson RD, Timmers PHA (1995) Rehabilitation Robotics in Europe. IEEE Transactions on Rehabilitation Engineering. 3. 35-45.