AN INTEGRATION IMPLEMENTATION USING AN M3S INFRASTRUCTURE WITH CONSIDERATIONS FOR USERS, PRESCRIBERS AND DEVELOPERS

Andre Wisaksana1,3, Geb Verburg3, Stephen Naumann2,3 1Department of Mechanical Engineering, 2Institute of Biomedical Engineering, University of Toronto 3Rehabilitation Engineering Department, The Hugh MacMillan Rehabilitation Centre

ABSTRACT

There are a wide range of technical aids available to persons with disabilities; these devices help them in their activities of daily living. This paper examines the applicability and appropriateness of the proposed rehabilitation communications standard, M3S, both in concept and implementation, in the context of three former clients with high level spinal cord injury who are users of technology. This paper also describes an M3S system implementation presently underway, and concludes by proposing one way in which M3S may be promoted, disseminated, and (ultimately) incorporated by rehabilitation developers, manufacturers and prescribers in North America.

BACKGROUND

Many assistive devices are available to persons with disabilities. These devices range from powered mobility, augmentative and alternative communications (AAC), environmental control, vocational tools (e.g. office equipment), to other more custom-devices (e.g. robotic manipulator). These technical aids have significant benefits for persons with disabilities in their daily living activities. We have approached three former clients of The Hugh MacMillan Rehabilitation Centre with regards to the types of practical assistive technologies they presently use in their home and work environments.

All three have sustained high level spinal cord injuries, and are frequent users of assistive technologies. Their typical system configuration consists of a powered wheelchair platform with interfaces to an onboard environmental control unit and a commercial home automation system; all three clients are also users of a computer oriented workspace. One of the clients has tested and used the Myoarm, a wheelchair-mounted robotic manipulator developed at The Hugh MacMillan Rehabilitation Centre [1].

STATEMENT OF PROBLEM

There are three main environments in which assistive technologies play a role. These are: (i) the wheelchair environment; (ii) the home environment; and (iii) the workspace environment.

Wheelchair environments consist of a commercial powered wheelchair platform and a corresponding input device (e.g. sip-and-puff, ultrasonic head switch); the wheelchair provides mobility and allows for physical access to other parts of the environment (workspace, home and community environment). Furthermore, due to restricted ability to access and control other types of devices in the other environments, the wheelchair input device typically incorporates some of the control modalities of the other two environments. Within the home environment, the functions typically accessed include: telephone and answering machine, entertainment systems (e.g. TV, VCR), lighting control, and door control; these functions are accomplished either through on-board (powered chair) or fixed location control (e.g. separate sip-and-puff). This functionality provides a measure of independence in daily living. It should also be noted that the bed environment is a subset of the home environment, and represents an important control location for the user outside of the wheelchair. The workspace environment includes the computer, monitor, printer, fax and related technologies, all accessed through each individual's control devices.

The persons with high-level spinal cord injury whom we have informally surveyed, and others with regressive conditions, often have difficulty operating devices, and yet they must be able to reliably control more devices than most people. Our clients' current systems are composed of combinations of different technologies performing different tasks or functions, each operated by their own input and control methods. Entertainment systems are typically controlled via IR switches and scanners, wheelchairs are controlled by direct switches (sip-and-puff), lights by X10 or remote X10, and telephones by other input methods.

Most of the access efforts to date have focused on the development and refinement of specific classes / brands of devices / appliances. The integration of control across the environments typically incorporates an ad hoc approach and custom hardware interfacing; this is due to the need to tailor access methods to the individual client on the one hand, and to the lack of an infrastructure around which to build the system on the other. As a consequence, if a user's ability level changes (e.g. from being able to control a four switch joystick to a sip-and-puff control) or a new function is needed, the system must be reconstructed. The encumbrance of modifying an existing control system often exceeds the available resources due to the diversity of systems, their proprietary nature, and need to revise the existing control structure. The latter case was described by two of the clients.

PROPOSED SOLUTION

A task-based system addresses the issue of integration of control. Such a system would be built around the tasks to be accomplished, which are tailored to the individual client's needs [3]. The main obstacle to its implementation is the large interfacing effort required, and the lack of a communications infrastructure for the modules [4]. An open rehabilitation communications standard would allow devices to be developed independently and maintain interoperability; furthermore, a standard would provide a well-defined means for interfacing device interactions to the home and workspace environments. This standard has been proposed in Europe by an international consortium, and is known as Multiple-Master Multiple-Slave (M3S) [5].

OVERVIEW OF THE M3S PREMISE

M3S is a proposed open communications standard for the rehabilitation environment. The M3S protocol defines the physical communications medium (CAN bus), as well as the message format of the information to be exchanged. There is presently no non-proprietary North American counterpart to the proposed M3S standard.

M3S operates as a task-based system in which tasks that need to be performed are defined from the various device modules attached to the bus; these tasks are defined in the system configuration software as opposed to being hard-wired. In order to achieve this, M3S makes use of the concept of degrees-of-freedom (DOF), which refers to the number of independent control parameters of a device. Degrees-of-freedom may have descriptive groupings (e.g. digital DOF for individual switches, analog DOF for joysticks); they may also be grouped by use, i.e. input degree-of-freedom (IDOF), or output degree-of-freedom (ODOF). A typical joystick has two degrees-of-freedom because it has two independent control parameters (Forward and Back as one, and Left and Right as the other). With this concept persons and devices may be described in terms of available and required DOFs respectively. The M3S system has a Configuration and Control Module (CCM) which is a program used to match the available inputs to the necessary outputs.

M3S SYSTEM IMPLEMENTATION

A project is presently underway at The Hugh MacMillan Rehabilitation Centre. The project goals are to develop a communications infrastructure for integration of control using a task-based control structure, and to develop a better means of interacting with the other environments (home and workspace). The tasks to be performed determine the number and type of modules that are to be incorporated into the system. There are two stages to the project: the first stage involves the implementation of M3S on a commercial (Invacare) powered chair platform. One of the M3S modules to be developed includes an RF remote link. The purpose of the link is to allow for transparent control of other systems without the constraint of being tethered to a fixed location. The second stage involves the development of a means to interact with a commercial home control system (CEBus). Development efforts are presently underway.

The M3S system requires the use of embedded control. This allows both flexibility of the overall system, and simplifies any subsequent modifications or expansions that need to be made. The tradeoff is that the system is technically more involved (i.e. use of embedded control). From the user and prescriber's perspective, this is additional involvement is not noticeable, and the main advantage is flexible configuration of the task tree, allowing the control structure to be more easily tailored and optimized to the user's needs. For the technologist/engineer implementing M3S modules, development of device application programs are greatly simplified as system knowledge is not necessary.

DISCUSSION

The M3S premise is both applicable and appropriate in the context of our work with our three former clients. The implementation presently under development represents a new approach in that a task-based system tailored to the individual user is built around a defined infrastructure; this defined infrastructure allows extensions and modifications to the user's system as necessary. Furthermore, the M3S system appears to provide a coherent means of accessing the other environments, such as home automation systems (e.g. CEBus). There are other potential applications for M3S, which may include roles as a tool for assessing different powered wheelchairs and control strategies. The flexibility in M3S is also applicable to control a sensory stimulation environment or a snoezelen room. Thus, further exploration of the M3S concept is appropriate.

In order to disseminate the idea of M3S, and to explore other potential applications, our group is pursuing the formation of a North American (NA) counterpart to the European M3S Consortium, but with a slightly different function. Aside from maintaining and refining the protocol, the main role of the NA-M3S Consortium would be to promote and disseminate expertise and information. A consortium offers a cost-effective approach, and allows for cost-sharing. It also allows for input and feedback from a broad spectrum of developers, manufacturers and prescribers.

REFERENCES

[1] G Bush, I. Al-Temen, J. Hancock, J. Bishop, E.M. Slack, I. Kurtz, "Development of a Myoelectrically Controlled Wheelchair Mounted Object Manipulator for Quadriplegics", Proceedings of the Fourth ICORR, Wilmington, DE, 1994, 103-6.

[2] P.J. Guerette & R.J. Nakai, "Case Studies of Performance Using Integrated and Distributed Controls", Proceedings 17th RESNA Conference, Nasvhille, TN, 1994, 287- 9.

[3] F. Shein, "Matching Tasks, Devices and User Abiilities for Selecting Controls for Access and Powered Mobility", Proceedings of CSUN '95,Los Angeles, CA, 1995.

[4] G. Verburg, M. Milner, S. Naumann, J. Bishop, O. Sas, "An Evaluation of the MANUS Wheelchair-mounted Manipulator", Proceedings 16th RESNA Conference, Toronto, ON, 1992, 602-4.

[5] K. van Woerden, " A Safe and Easy to use Integrated Control and Communication Method", Rehabilitation Technology, Strategies for the European Union, IOS Press,1993, pp 75-80.

NOTE

Devices used include TOSC, CHEC I/II, Tykriphone, Peachtree controller, Unidialler, X10 controller and modules.

ACKNOWLEDGMENTS

The authors would like to acknowledge The National Strategy for the Integration of Persons with Disabilities, Industry Canada, and the Ontario Rehabilitation Technology Consortium for their financial support.

We would also like to thank Jack McLellan, Lew Boles and Steve McPherson for their valuable contributions to this paper.

Contact: Andre Wisaksana 350 Rumsey Road Toronto, ON M4G 1R8 CANADA phn: (416) 424-3855 ext. 734 fax: (416) 425-1634 wisaka@me.toronto.edu