Assistive Technology in Neuromuscular Diseases

1
Assistive Technology in Neuromuscular Diseases

• -F. Corno, L.Farinetti, I. Signorile, A
  Cost-Effective Solution for Eye-Gaze Assistive
  Technology.
• -R. Cooper, Intelligent Control of Power
  Wheelchairs.

Mike Wininger Rutgers
BME 27september2004
2
What is Assistive Technology (AT)?

• Device designed to to maintain or improve the
  functional capabilities of a person with a
  disability.

-Examples
Augmentative communication devices,
computers,
powered mobility,
wheelchairs
3
What is the patient population?

• Severe arthritis patients
• Stroke victims and sufferers of various degrees
  of paralysis
• Any individual with a compromised mobility,
  communication, or dexterity

Sufferers from Neuromuscular diseases (Muscular
Dystrophy MD, Motor Neuron Disease MND/Lou
Gehrigs Disease/ALS, Parkinsons Disease, etc)
4
Priorities in Daily Living

• Communication and articulation
• Hygiene
• Completion of simple daily tasks
• Mobility

From D.P. Romilly, et.al 1994.
5
What do we have to work with?

• Most NMD patients have some use of extremities
• Nearly all patients remain unimpaired in
  intellect/learning ability
• Most have use of all the afferent senses, notably
  touch and sight, (note that ocular motor cortex
  is often unaffected by NMDs).

6
A Cost-Effective Solution for Eye-Gaze Assistive
Technology- Fulvio Corno, L.Farinetti, I.
Signorile.

• Associate Professor at Politecnico di Torino,
  Italy
• Web-Intelligence, digital systems, genetic
  research
• Proposes affordable means of communication for
  ALS/MND patients.

7
Background

• Communication is a high priority for the social
  species H. sapiens
• Nerve deterioration prevents NMD patients from
  communicating at their desired level
• Eye-Gaze technology exists and has proven itself
  to be an effective means of learning and
  communicating, with some drawbacks

8
Eye-Gaze Technology

• Gaze is an intuitive and primal action which can
  be a useful tool for communication
• A pupil-tracker is a way of calculating the
  position of an individuals gaze and can be
  incorporated with an image-processor and display
  device to create a communication-by-selection
  process
• Most eye-gaze devices are bulky and intrusive,
  compromising a patients preference for discreet
  assistance or expensive.

9
Gaze Computation

• Computer screen displays
• images in each of 6 regions
• Gaze location is calculated
• to be in one of these regions
• and that character is chosen

• A momentary gaze (a glance) is given as any gaze
  in which the first
• calculated visual acquisition is not followed
  immediately by a second.
• That is, a minimum threshold of three gaze
  samples is requisite for a
• character to be chosen.

10
Acquisition Process
A (black, not yellow) marker on the forehead
serves as reference while the camera and image
processing program searches for the dense
concentration of black at the pupils for eye
location and gaze calculation
11
Acquisition Process
12
Benefits of Gaze Computation

• Beneficial for high-level quadriplegia, NMD,
  brain-stem stroke sufferers.
• Most families have a computer and can purchase a
  web cam, and the software may be made shareware.
• No bulky apparatus, little encroachment on
  personal space, little assistance needed.
• Possible merger to other applications- improved
  resolution may create a means for web exploration
  and household conduction.

13
Problems with Gaze Computation

• Artifact Contact lenses may interfere with
  signal tracking, temporary distraction may
  produce error
• Survey time
• Limited resolution
• Eye strain and fatigue

14
Quandaries

• In the Corno paper, fixation is computed when a
  gaze is held in a region for 2-3 seconds. This
  allows for fairly rapid user icon acquisition and
  allows a thought to be communicated in
  short-order.
• However, if the user has trouble seeing and needs
  to focus, loses focus or concentration, or
  perhaps wants to repeat an icon, the sample limit
  will not effect the desired communication.

15
Image acquisition

• The focus process (especially for special
  populations) can last from 10s of milliseconds
  to seconds.
• A vergence of 4 degrees requires between .5 and
  .75 seconds to allow for image acquisition.
• Reading (especially for youngsters) or image
  contemplation may require seconds.

Semmlow, JL, and Yuan, W 2002
16
Eye strain and Fatigue

• Muscular endurance is compromised in NMDs.
  Initial strength of ocular muscles may be within
  a normal range, but repetition breeds continual
  decrease in performance until total fatigue
  occurs.

17
Intelligent Control of Power Wheelchair -
Rory A. Cooper

• Chair of Rehabilitation Science and Technology at
  the University of Pittsburgh
• Wheelchair and Assistive Technology
• Discusses navigation and safety mechanisms of
  intelligent wheelchair technology.

18
Background

• Mobility is the exercising of independence and
  demonstration of self-sufficiency
• Motor Neuron Diseases are characterised by loss
  of dexterity, gait impairment, loss of limb
  movement, and often paralysis.
• Intelligent wheelchairs provide a means for
  comfortable navigation and functional mobility
  while accommodating an individuals limitations.

19
User Interface

• Must be complex enough to manage the needs of the
  user, but be simple enough to ensure ease of
  operation and minimize the risk of failure.
• Interface selection must be matched by user need
  and abilities.

20
Joystick

• Simple control device
• Easy to manufacture, install, and troubleshoot
• Highly reliable
• Intuitive manipulation
• Can be employed in many wheelchairs

Joysticks with varying degrees of complexity
21
Sip/Puff

• For users with highly restricted mobility, this
  is an effective means of translating mechanical
  input to electrical output.

22
Key Pad

• For users with dexterity, this maximizes the
  range of user options and allows for integration
  of multiple assistive devices.

23
Quandaries.

• If a wheelchair user has advanced muscular
  atrophy in the upper limb (and/or) compromised
  dexterity, digital manipulation of controls is
  difficult, and so the user interface must have a
  light buffer so that it is easy to operate and
  maintain control of the chair
• However, patients with Motor Neuron Diseases can
  often have tremor or insufficient gross limb
  articulation, resulting in sometimes inaccurate
  finger placement on a fine-touch control board,
  which will lead to mis-touches and unintentional
  wheelchair movement

24
Joystick signal filtering

• Pros
• Unintentional input as a result of tremor or
  momentary loss of coordination will be dampened,
  creating a smoother ride
• Filter sensitivity can be set to accommodate
  users ability
• Cons
• Fine adjustments in direction or speed are more
  difficult to enact
• Creation of dead zone around mean joystick
  position

25
1- and 2-filter damping
Simple white noise input with 1 and 2 low-pass
discrete filters. Resulting magnitudes are damped
by 60 and 90, respectively.
26
2-filter tremor model
Figure Moving a joystick 4 units in one
direction with tremor of a small-medium deviation
about this mean position .
27

• Medium tremor about a mean joystick displacement
  filtered through 2 low-pass filters condenses
  signal about a new mean, displaced slightly from
  the intended output.

28
Transfer Function filter

• This filter reduces jerk at initial input, with
  equal displacement thereafter.

29
Fine joystick control

• In slow, ramping control of the joystick, the
  transfer function filter is very precise,
  particularly in the initial displacements.

30
Shared Control

• Goal to enhance mobility of the user by
  off-loading some wheelchair driving functions to
  an automatic control system.
• Some users have sensory, cognitive, and/or
  physical impairments that limit their ability to
  control a power wheelchair.
• Shared control mechanisms are regular features in
  our lives power steering, computerized exercise
  equipment, stereo equalizers...

31
Quandaries.

• If a wheelchairs navigation control is set to
  avoid obstacles that come within a certain radius
• However if a user wants to navigate through a
  doorway or dock at a desk, the control mechanism
  would prevent this action

32
Quandaries. (Shared Control)
A shared control chair has two options to avoid
a perceived obstruction or to allow user input to
intercede in the instruction loop.
Docking and door passage present two conflicting
operations for autonomous control devices.
33
NaviChair Shared Control design

• Laser range finders, sonar, and ultrasound
  sensors are the most commonly used sensor devices.

• NaviChair works on sonar principle signal is
  sent from the chair,

and
the time until reception dictates object
distance.
34
NaviChair Shared Control

• Two settings High Autonomy (chair-dependent) and
  Low Autonomy (user-dependent).
• High Autonomy User input has no change in chair
  navigation.
• Low Autonomy Operates on same obstacle avoidance
  principle, but can be attenuated or over-ridden
  by touch of the joystick.
• Does not guarantee that chair will move in an
  obstacle free path, so a collision-prevention
  routine slows the chair by an amount proportional
  to the square root of the distance to the nearest
  obstacle in the direction of motion.
• Control modes can be changed by adjusting the
  autonomy of the obstacle avoidance.

35
NaviChair Shared Control

• Pros
• Variability of autonomous control
• Collision-prevention maintains safe distance from
  obstacles with slow approach
• Program learns the users behavior and
  cooperates with the user.

• Cons
• Expense
• Autonomy control interface
• Diffuse reflection?
• Oddly-faceted objects?

36
Citations

• -D.P. Romilly, et.al A Functional Task Analysis
  and Motion Simulation for the Development of a
  Powered Upper-Limb Prosthesis. IEEE Transactions
  on Rehabilitation Engineering, Vol 2. No 3,
  September 1994
• -F. Corno, L.Farinetti, I. Signorile, A
  Cost-Effective Solution for Eye-Gaze Assistive
  Technology.
• -Assistive Technologies, Principles and Practice,
  2/e Cook, Hussey. Mosby 2002.
• -Semmlow, JL, and Yuan, W, Adaptive Modification
  of Disparity Vergence Components An Independent
  Component Analysis study. Investigative
  Ophthalmology and Vision Science (In Press).
• R. Cooper, Intelligent Control of Power
  Wheelchairs.
• T. Röfer, A. Lankenau (1999). Ensuring Safe
  Obstacle Avoidance in a Shared-Control System. In
  J. M. Fuertes (Hrsg.), Proc. of the 7th Int.
  Conf. on Emergent Technologies and Factory
  Automation, S. 14051414.
• Levine, S., Koren, Y., Borenstein, J. (1990).
  NavChair Control System for Automatic Assistive
  Wheelchair Navigation. In the Proceedings of the
  13th Annual RESNA International Conference.
  Washington, D.C. RESNA, 193-194.