Infant Control of a Mobile Robot more

RESNA  Annual  Conference  –  June  26  –  30,  2010  –  Las  Vegas,  Nevada Making Assistive Technology and Rehabilitation Engineering a Sure Bet Infants Control of a Robotic Mobility Device Madeline E. Smith, Carole Dennis, Sc.D., OTR, Sharon Stansfield, PhD, Hélène Larin, PT, PhD Ithaca College Ithaca, NY 14850 ABSTRACT Independent  mobility  is  crucial  in  the  development  of  typical  infants.    Children  with  physical  disabili9es   may  experience  concomitant  emo9onal  and  psychological  limita9ons  a<ributable  to  restric9on  of   movement.    Robots  with  sonar  have  been  suggested  as  a  viable  means  of  providing  mobility  for  infants   with  physical  disabili9es.    We  have  developed  such  a  system  using  a  Pioneer  3  robot  and  various  input   devices.    In  our  system  the  infant  is  posi9oned  in  a  commercial  child  seat  on  top  of  the  robot  and   allowed  to  control  the  robot’s  movement  with  various  control  methods.    This  paper  describes  one   control  method  for  our  system  that  uses  a  Wii  balance  board  to  detect  when  the  infant  leans  and  then   moves  the  robot  in  the  direc9on  of  the  lean.    We  use  parameter  files  to  customize  the  system  for  each   child.     KEYWORDS Infant,  Mobility,  Robot,  Balance  Board,  Posture  Tracking BACKGROUND The  first  independent  mobility  of  the  human  being  is  achieved  with  the  development  of  crawling  and   walking,  which  occurs  in  typically  developing  infants  between  six  and  14  months  of  age.    Independent   mobility  is  crucial  in  the  development  of  typical  infants,  as  it  allows  them  to  acquire  a  broad  range  of   skills  in  the  cogni9ve,  perceptual,  emo9onal,  and  social  domains  [(1),  (2),  (3)].    Typically  developing   children  whose  mobility  has  been  restricted  for  even  rela9vely  short  periods  of  9me  have  been  found  to   demonstrate  apathe9c  behavior  and  depressed  mo9va9on  (4).    Children  with  physical  disabili9es  and   restricted  mobility  have  been  found  to  demonstrate  increased  dependence,  frustra9on,  depressed   mo9va9on,  lack  of  curiosity,  and  a  lack  of  confidence  (4).     Several  inves9gators  have  reported  that  children  with  physical  disabili9es  may  achieve  significant   developmental  benefits  from  the  use  of  powered  mobility,  including  increased  interac9on  with  the   environment  and  contact  with  others,  self-­‐ini9ated  independent  movement,  posi9ve  affect,  and   communica9on  [(5),  (6)].    Powered  mobility  is  costly,  imposes  safety  risks,  and  tradi9onally  has  not  been   recommended  un9l  children  are  24  to  36  months  of  age  [(7),  (8)].    Younger  children  may  also  benefit   from  the  developmental  opportuni9es  provided  through  the  use  of  powered  mobility.    However,  very   limited  research  has  been  conducted  to  inves9gate  these  poten9al  benefits.    Galloway,  Rhu,  and  Agrawal   (8)  have  suggested  that  the  use  of  a  robot  may  be  a  viable  means  of  providing  mobility  for  infants  and   young  children  with  physical  disabili9es. PROJECT GOALS The  long-­‐term  goal  of  this  research  is  to  inves9gate  the  developmental  benefits  of  providing  safe   powered  mobility  to  infants  with  physical  disabili9es  who  are  less  than  two  years  of  age.    Our  immediate   Copyright © 2010 RESNA 1700 N. Moore St., Suite 1540, Arlington, VA 22209-1903 Phone: (703) 524-6686 - Fax: (703) 524-6630 1 RESNA  Annual  Conference  –  June  26  –  30,  2010  –  Las  Vegas,  Nevada Making Assistive Technology and Rehabilitation Engineering a Sure Bet goal  is  to  develop  a  custom  system  allowing  pre-­‐mobile   infants  to  display  goal-­‐directed  movement.    We  will  be   running  a  pilot  study  in  early  2010  to  determine  an   op9mal  control  method  for  our  robo9cs  system  that  will   allow  typically  developing  infants  to  demonstrate   purposeful  mobility  within  a  structured  environment.    We   hope  to  work  with  approximately  15  typical  infants   between  the  ages  of  4  and  12  months  of  age.    The  infants   will  provide  informa9on  about  which  systems  can  be   controlled  by  the  infant  at  an  early  age;  that  informa9on   will  allow  us  to  calibrate  the  soeware  systems  to  allow   babies  of  different  weights,  lengths,  and  motor  abili9es  to   efficiently  control  the  robot.    Following  tes9ng  with   typically  developing  infants,  we  expect  to  determine  the   Photo 1: Robot best  op9on  for  control  of  the  robot,  develop  an  infant   training  program  to  use  the  device,  and  then  to  assess   developmental  benefits  that  may  be  associated  with  the  use  of  the  robot. ROBOT SYSTEM The  robot  system  we  are  using  consists  of  a  Pioneer  3  robot,  a  custom  sea9ng  plahorm,  a  commercial   child’s  seat,  commercial  input  devices,  and  custom  robo9cs  soeware.    The  Pioneer  3  robot  is   commercially  available  and  is  equipped  with  sonar  which  we  use  to  detect  obstacles  and  prevent   collisions.    Photo  1  shows  the  robot  and  its  sonar  sensors.     We  designed  and  built  the  custom  plahorm  to  fit  over  the   robot  and  support  the  weight  of  the  child,  seat  and  input   devices.    A  commercial  wireless  joys9ck  acts  as  a  master   override  switch  and  allows  the  therapist  to  remotely   control  the  robot  and  to  stop  movement  at  any  9me  for   safety  reasons.    The  other  commercial  input  devices  we  are   using  include  a  gaming  joys9ck,  a  child’s  joys9ck  and  a  Wii   Fit  Balance  Board.    We  have  developed  the  robo9cs   soeware  to  run  on  the  robot’s  on-­‐board  computer,  allowing   an  infant  to  control  movement  using  one  of  these  input   devices.    Our  soeware  monitors  the  sonar  to  prevent   collisions  while  gathering  robot  mo9on  data  from  the  robot   wheel  encoders.     BALANCE BOARD CONTROL The  most  novel  of  our  control  methods  makes  use  of  the   balance  board.    This  device  is  designed  for  use  with  the   Nintendo’s  Wii  gaming  system  for  fitness  games  and   Figure 2: Balance Board Control Copyright © 2010 RESNA 1700 N. Moore St., Suite 1540, Arlington, VA 22209-1903 Phone: (703) 524-6686 - Fax: (703) 524-6630 2 RESNA  Annual  Conference  –  June  26  –  30,  2010  –  Las  Vegas,  Nevada Making Assistive Technology and Rehabilitation Engineering a Sure Bet exercises.    For  our  robot  system,  the  child  sits  in  the  commercial  seat  on  top  of  the  balance  board,  which   lies  on  the  custom  plahorm.     The  balance  board  has  responsive  pressure  sensors  in  each  of  its  four  corners  and  built  in  Bluetooth   capabili9es.    We  use  a  commercial  Bluetooth  adapter  and  WiiYourself!  C++  library  (9)  to  access  balance   board  data  within  our  robo9cs  soeware.    Our  soeware  compares  the  values  of  the  four  pressure  sensors   to  determine  which,  if  any,  direc9on  the  infant  is  leaning.    When  a  sustained  lean  is  detected  the  robot   begins  to  move  in  that  direc9on.    When  the  child  reaches  out  for  an  item,  he  or  she  leans  in  the  direc9on   of  the  item  and  moves  toward  it.    We  believe  this  will  be  the  most  intui9ve  method  for  a  young  child  to   learn  to  use.     CUSTOMIZATION Each  child  moves  in  different  ways.    To  account  for  this  our  robo9cs  soeware  has  a  number  of   parameters  to  adjust  opera9on.  The  balance  board  method  uses  four  threshold  values  to  determine  the   percentage  of  a  child’s  weight  that  must  be  on  one  side  to  cons9tute  a  lean.    Other  parameter  values   include  the  speed  at  which  the  robot  should  move  and  the  stopping  distance  to  avoid  collisions.    We   create  a  file  to  store  these  values  for  each  child.    The  first  9me  a  child  uses  the  system  ,the  therapist  can   input  parameter  values,  use  default  values,  or  use  calibrated  values.    During  the  calibra9on  the  child  is   prompted  to  lean  in  each  of  the  four  direc9ons.    Readings  are  taken  during  these  leans  to  compute  the   threshold  values.    The  parameter  values  can  be  easily  adjusted  at  any  9me  to  improve  system   performance  by  edi9ng  this  file.     OBSERVATIONS While  we  have  not  yet  performed  any  formal  studies,  we  have  had  three  infants,  aged  7  to  9  months,  use   the  system  as  we  developed  the  soeware.    All  of  the  infants  were  able  to  move  the  robot  as  they  leaned   to  get  an  offered  toy  or  drink.    Children  were  also  observed  to  move  the  robot  spontaneously  when  they   were  not  being  directly  offered  something  or  coaxed.    The  robot  was  responsive  to  the  infants'  upper   body  weight  shie  as  they  leaned,  moving  to  the  lee,  forward,  and  the  right  according  to  the  direc9on  of   the  child's  movement.    Although  it  is  not  possible  at  this  9me  to  determine  whether  the  children  realized   that  their  leaning  caused  the  robot  to  move,  we  an9cipate  that  this  will  become  apparent  as  children  are   offered  repeated  experiences  with  the  robot.     REFERENCES 1. Bertenthal,  B.L.,  Campos,  J.J.,  &  Barre<,  K.  (1984).    Self-­‐produced  locomo9on:    An  organizer  of   emo9onal,  cogni9ve,  and  social  developments  in  infancy.    In  Robert  N.  Emde  and  Robert  J.  Harmon   (Eds.),  Con9nuites  &  incon9nui9es  in  development,  pp.  175-­‐210.  .  New  York  and  London:  Plenun   Press 2. Bertenthal,  B.L.,  &  Campos,  J.J.  (1987).    New  direc9ons  in  the  study  of  early  experience.    Child   Development,  58,  560-­‐567.     3. Thelen,  E.  (2000).  Grounded  in  the  world:  developmental  origins  of  the  embodied  mind.  Infancy,  1,   3-­‐28. Copyright © 2010 RESNA 1700 N. Moore St., Suite 1540, Arlington, VA 22209-1903 Phone: (703) 524-6686 - Fax: (703) 524-6630 3 RESNA  Annual  Conference  –  June  26  –  30,  2010  –  Las  Vegas,  Nevada Making Assistive Technology and Rehabilitation Engineering a Sure Bet 4. Butler,  C.  (1986).    Effects  of  powered  mobility  on  self-­‐ini9ated  behaviors  of  very  young  children  with   locomotor  disability.    Developmental  Medicine  &  Child  Neurology,  28,  325-­‐332. 5. Deitz,  J.,  Swinth,  Y.,  &  White,  O.  (2002).    Powered  mobility  and  preschoolers  with  complex   developmental  delays.    The  American  Journal  of  Occupa9onal  Therapy,  56,  86-­‐96. 6. Nicholson,  J.,  &  Bonsall,  M.  (2002).  Powered  mobility  for  children  under  five  years  of  age  in  England.   Bri9sh  Journal  of  Occupa9onal  Therapy,  68,  291-­‐293. 7. Cox,  D.  L.  (2003).    Wheelchair  needs  for  children  and  young  people:    a  review.    Bri9sh  Journal  of   Occupa9onal  Therapy,  66(5),  219-­‐223. 8. Galloway,  J.  C.,  Ryu,  J.,  &  Agrawal,  S.K.  (2008).    Babies  driving  robots:    Independent  mobility  in  very   young  infants.    Journal  of  Intelligent  Service  Robo9cs,  1,  123-­‐134. 9. gl.<er  (2007).    Wii  Yourself!  Na9ve  C++  Wiimote  Library.    h<p://wiiyourself.gl.<er.org/ ACKNOWLEDGMENTS Keith  Taylor  (Washington  State  ’10)  and  Virginia  Nearing  (Saint  Mary’s  ’10)  also  worked  on  developing   this  robo9cs  soeware  as  part  of  a  summer  2009  Research  Experiences  for  Undergraduates  project  at   Ithaca  College  funded  by  the  Na9onal  Science  Founda9on.     Lauren  Cresser  (Ithaca  College  ’11)  is  assis9ng  with  this  project,  funded  by  an  Ithaca  College  DANA   internship.     Author Contact Information: Madeline  E.  Smith,  Ithaca  College,  Williams  Hall,  953  Danby  Rd,  Ithaca,  NY  14850,  PHONE:  (610)   730-­‐2837    EMAIL:  msmith11@ithaca.edu Copyright © 2010 RESNA 1700 N. Moore St., Suite 1540, Arlington, VA 22209-1903 Phone: (703) 524-6686 - Fax: (703) 524-6630 4
x

Log In

or reset password

Reset Password

Enter the email address you signed up with, and we'll send a reset password email to that address

Academia © 2012