Elaine Lu, University of Toronto
ABSTRACT
User centred design techniques were used to develop a rehabilitation robotic device that would increase access to rehabilitation and give therapists another tool in the rehabilitation of the upper limb of stroke survivors. A survey was conducted to evaluate therapists’ needs and desires in such a device. A prototype was then designed using the house of quality matrix that would be portable, easy to use, and cost effective. Focus groups will further hone the design, which will be pilot tested in the future.
BACKGROUND
Every year stroke affects 16.3 million people [1]. As a result, many hours are spent on post-stroke rehabilitation to improve function, reduce impairment, and enable stroke survivors to live more independently. Motor rehabilitation is often a large part of this, as motor impairment affects approximately 80% of stroke survivors and can result in serious disability [2]. As the upper limbs are necessary for many activities of daily living, it is important for the upper limb to be supported and rehabilitated.
There are several factors that may result in inadequate access to stroke rehabilitation. These may include an uneven distribution of resources, inadequate insurance coverage, or a lack of knowledge of the potential value of rehabilitation [3]. An uneven distribution of resources may be due to location, for example those in remote locations may have less access to trained therapists than those in urban centres. In addition, staff may be limited due to a shortage of qualified stroke therapists. These issues may limit a stroke survivor’s therapy in terms of frequency and duration, which would be especially important during the critical time window for recovery.
PROBLEM STATEMENT
The goal of this project is to design a rehabilitation device to increase access to rehabilitation and aid stroke therapists. Even though rehabilitation robotic devices have been on the market for many years, not many are used in clinics or hospitals. This may be due to cost, size, or usability issues, which need to be addressed before robotic devices are able to increase access to rehabilitation.
Increasing access to rehabilitation will enable some stroke survivors to recover more fully, enabling to live more independent lives.
METHODOLOGY
User centred design practices were used in order to design a device clinicians and stroke survivors would use. The first phase of the design was an international online survey sent out to stroke therapists to better understand therapist requirements and preferences in an upper limb rehabilitation robotic device [4]. Features polled included approaches to rehabilitation, movement patterns for the device, as well as levels of assistance and feedback, interfaces for therapists and/or stroke survivors, types of interactive activities, and accessibility (size, portability, and cost).
Results from the survey were then used to inform the design of a prototype rehabilitation robot to be built by an industrial partner. The technical requirements of the robot were prioritized using the House of Quality matrix [5]. Customer requirements were derived from the survey, using percentages of respondents’ agreement of statements or importance of desired features. The percentages were multiplied by ten to assign a weight to each requirement. In addition, written comments that pertained to the design of the device, were included in the customer requirements, although not surveyed, they were given a value of ‘1’ for the weight. Customer requirements were put on an affinity chart and grouped according to similar characteristics.
Technical requirements to accomplish the needs of a rehabilitation robotic device were put in a correlation matrix, indicating which requirements have negative or positive correlations with each other.
The planning matrix was done with the prototype and an estimate of how well the planned prototype would be able to meet customer requirements. A mark from 0 to 5 was given to each customer requirement. This was multiplied by the customer requirement weight to give an overall weight to each requirement.
Technical priorities were calculated with an interrelationship matrix of the technical specifications and the customer requirements, with a ‘3’ indicating strong correlation, ‘2’ medium correlation, and a ‘1’ weak correlation, if there was no correlation, no number was assigned. These numbers were then multiplied by the overall weight from the planning matrix and summed for each technical requirement. Technical specification targets were created using anthropometric measurements and other data.
RESULTS
Customer Requirements
Out of 320 therapists who started the survey, 233 surveys were completed and analyzed. Categories of customer requirements were determined using an affinity chart. Five main categories were assigned: quality of rehabilitation, usability, accessibility, safety, and motivational factors. In terms of costs, 81% of therapists believed the robotic device should be below $10,000 USD.
Technical Requirements
Technical requirements were created in conjunction with the industrial partner. The technical requirements and the correlation matrix can be seen in Figure 1. The positive correlations were marked with a ‘++’ indicating strong positive correlation or a ‘+’ indicating moderate positive correlation. The negative correlations were similarly marked with ‘–‘ or ‘-‘ indicating strong or moderate negative correlation. Arrows were given to each requirement to indicate the direction (increase or decrease) the specification needed to go in for a better device.
Figure 1. Technical Requirement interaction matrix of the rehabilitation robotic device
Targets
Target specifications are given in Table 1, which also gives the priority of each target according to the interaction matrix (not shown) and, the percentage out of the sum of the total points for technical priorities. Targets were then generated using the 95th percentile of anthropometric data, normal values for range of motion, values for strength in rehabilitation, portability with regards to airplane carry on sizes and safe lifting standards [6-10]. Numbers of links and joints were calculated using a 2 degree of freedom model in order to decrease costs. Compatibility issues were addressed using current standards (USB plug-in, AC adaptor).
Table 1. Technical specifications and targets based on the house of quality matrix.
Prototype
The prototype was designed using SolidWorks 2010, see Figure 2 . In consultation with the industrial partner, compromises were made due to cost constraints. The resulting prototype dimensions were 457mm x 305 mm x 203 mm, weighing approximately 7.5 kg. The maximum workspace was 1016 mm x 298 mm x 762 mm, with a “sweet spot” of 254 mm x 457 mm. The continuous force capability at the home position was 13.2 N per plane of motion, and the maximum force capability was 52.8 N per plane of motion. The resolution at the home position was 0.0131 mm/count. The device used two DC motors and two optical encoders. It was a haptic device, able to sense user forces in order to simulate movements used in activities of daily living. It used a USB and AC connections as well as an emergency stop button. Although it was a 2 degree of freedom device, it can be used in several orientations, depending on which plane it would need to accommodate. Figure 4 shows the unfinished state of the device as it undergoes testing.
Figure 2. Conceptual design of the rehabilitation robotic device; overview of the interior
Figure 3. Conceptual design of the rehabilitation robotic device in the vertical position
Figure 4. Unfinished rehabilitation assistive device being tested
DISCUSSION
Designing a device which would meet the user requirements, and yet be cost effective was a challenge. In order to keep costs down, a modular unit two degree of freedom unit was proposed, one that was compatible with current technology, such has personal computers. Separate modules could be sold with the unit depending on user needs, for example a separate biofeedback unit. Module handholds would accommodate different users on the machine.
An end effector unit (as opposed to an exoskeleton) was chosen to allow therapists to have more interaction with their client. This would allow for more types of therapy to be performed on the stroke survivor. For example, the unit would allow for cueing by the therapists. In addition, bilateral therapy could be easily incorporated using two units.
Therapists desired a device that would be compact and usable in either clinic or home. In order to increase portability, and therefore accessibility to rehabilitation, a desktop model was proposed. The desktop model would allow for different positions depending how it was positioned. It would be able to be used in seated, standing, and even recumbent positions.
Although a two degree of freedom model does limit motion, this unit is able to be used in several configurations to accommodate different arm movements.
The motors and sensors on the device allow it to create haptic feedback, which along with visual and audio feedback would allow for more sensory rehabilitation, as well as be more true to activities of daily living. Haptic feedback would also be able to give resistance based on user performance.
IMPLICATIONS
The current cost of the prototype was $4120 USD per device to make two devices. This cost included an aluminum casing and machining of custom parts. The cost may be brought down if the casing is made from plastic and if the device is mass produced. After more refinement of the design through focus groups, the device will be manufactured on a small scale to enable clinical testing. As current actuated devices on the market sell for over $50,000 USD, this device would potentially increase access to robotic rehabilitation.
FUTURE WORK
Future work would include focus groups and pilot studies with stroke survivors and therapists to evaluate the usability of the device, as well as clinical trials to evaluate the effectiveness of the robotic therapy. Future directions include using artificial intelligence to adapt to the user. Different software packages that would be able to address different therapy approaches as well as motivate users would also need to be developed.
ACKNOWLEDGEMENTS
I would like to thank Alex Mihailidis, Rosalie Wang, Debbie Hebert, and Jennifer Boger, at the University of Toronto for their insight and direction.
Special thanks to those at our industrial partner, Quanser — Paul Karam, Don Gardner, Fayez Khan, Derry Crymble, Herve Lacheray at Quanser for technical advice and fabrication of the device.
REFERENCES
[1] T. Truelsen and R. Bonita, “The worldwide burden of stroke: current status and future projections,” Handbook of Clinical Neurology Vol 92 (3rd Series): Stroke Part I: Basic and Epidemiological Aspects, M. Fisher, ed., Amsterdam: Elsevier, 2009, pp. 327-336.
[2] P. Langhorne, F. Coupar, and A. Pollock, “Motor recovery after stroke: a systematic review,” Lancet Neurology, vol. 8, 2009, pp. 741-754.
[3] G.E. Gresham, D. Alexander, D.S. Bishop, C. Giuliani, G. Goldberg, A. Holland, M. Kelly-Hayes, R.T. Linn, E.J. Roth, W.B. Stason, and C.A. Trombly, “Rehabilitation,” Stroke, vol. 28, 1997, pp. 1522-1526.
[4] E. Lu, R. Wang, D. Hebert, J. Boger, M. Galea, and A. Mihailidis, “The development of an upper limb stroke rehabilitation robot: Identification of clinical practices and design requirements through a survey of therapists,” Disability and Rehabilitation: Assistive Technology, vol. In Press, 2010.
[5] J.R. Hauser and D. Clausing, “The house of quality,” Harvard business review, vol. 66, 1988, p. 63–73.
[6] Ergonomic guidelines for manual handling, 2nd ed., WorksafeNB, St. John, NB, 2010.
[7] Air Canada. (2009). Aircanada – Carry-on Baggage [Online]. Available: http://www.aircanada.com/en/travelinfo/airport/baggage/carry-on.html
[8] S.B. Thies, P.A. Tresadern, L.P. Kenney, J. Smith, D. Howard, J.Y. Goulermas, C. Smith, and J. Rigby, “Movement variability in stroke patients and controls performing two upper limb functional tasks: a new assessment methodology.,” Journal of neuroengineering and rehabilitation, vol. 6, Jan. 2009.
[9] M.H.M. Lee and A. Moroz (2009), “Physical Therapy,” in The Merck Manuals Online Medical Library for Healthcare Professionals and Caregivers [Online]. Available: http://www.merckmanuals.com/professional/sec22/ch336/ch336b.html?qt=rehabilitation&alt=sh.
[10] R.Q. Van Der Linde, P. Lammertse, E. Frederiksen, and B. Ruiter, “The HapticMaster , a new high-performance haptic interface,” FCS Control Systems, 2001, pp. 1-5.
Author correspondence: Elaine Lu, 54 Longford Crescent, Toronto, ON M1W1P4, Phone: 416-498-9285