Dimitar Stefanov

1
Lecture 25

• Dimitar Stefanov

2
Autonomous-Guided Wheelchairs
Go-to-goal wheelchairs
3
Wheelchair Control, based on Visual servoing of
the head position
Shirai Lab (1997-1998)
4
Wheelchair Control, based on Visual servoing of
the head position (continue)
5
Wheelchair Control, based on Visual servoing of
the head position (continue)
6
TAO-1 Intelligent WheelchairApplied AI Systems
Inc.
7
TAO-1 Intelligent WheelchairMain characteristics

• Infared and bump sensors
• Automatic visual avoidance
• Voice command response
• Collision avoidance
• Passage through a narrow corridor
• Entry through a narrow doorway
• Landmark based navigation

8
TinMan intelligent wheelchair controllerMain
characteristics

• KISS Institute for Practical Robotics (KIPR)
• supplemental wheelchair controller that can be
  retrofitted to existing wheelchairs
• safely and independently operation a powered
  wheelchair by users who has partial visual
  impairment or brain damage,
• sensors for obstacles detection

9
TinMan intelligent wheelchair controller
(continue)
10
Light guidance system Dohi Lab
11
Autonomous guided wheelchair Nagasaki
University and Ube Technical College

• position error 0.35 m
• angular error 17 degrees

• uses existing ceiling lights
• vision sensor (position)
• azimuth sensor (orientation)
• wheels angle rotation sensor (odometric
  information)
• laser range sensor (obstacles detection)

12
MAid project Research Institute for Applied
Knowledge Processing FAW

• robotic wheelchair for transport of elderly
• semi-autonomous mode
• autonomous mode
• wheel encoders
• fiber-optic gyroscope
• sonar system
• infrared sensors SICK

13
Wheelesley Intelligent wheelchair
14
Wheelesley(continue)

• started at Wellesley College in 1995 (Holly
  Yanco)
• Developed at the KISS Institute
• moved to the MIT Artificial Intelligence
  Laboratory
• interface EagleEyes system (EOG -
  electro-oculographic potential)

15
Wheelesley EagleEyes system
16
NavChairUniversity of Michigan
17
NavChair(continue)

• University of Michigan (Simon Levine, Johann
  Borenstein)
• obstacle avoidance, follow walls
• narrow doorway passage

18
NavBelt University of Michigan
Device for guidance of blind people.
NavBelt generates acoustic cues conveyed to the
user via headphones.
19
GuideCane University of Michigan
Device for guidance of blind people.

• Fully automatic ultrasonic sensor-based obstacle
  avoidance
• Position information by combining odometry,
  compass, and gyroscope data

20
Drive Assistant (cont)
21
Drive Assistant (continue)

• VTT Machine AutomationTampere, Finland
• vehicle positioning and navigation
• dead reckoning
• differential GPS
• passive transponders
• natural landmarks in the environment
• laser based navigation
• part of the project FOCUS for the TIDE programme
• ultrasonic sensors
• M3S interface.

22
SENARIO (1994)
23
SENARIO(Ultrasonic sensors)
24
SENARIO(continue)

• Intelligence in the navigation systems of the
  powered wheelchair
• Autonomous mode - "go to goal" commands
• Obstacles and risks avoidance system.

25
Intelligent wheelchair at the University of Notre
Dame (1994)
26
PAM - AID projectPersonal Adaptive Mobility Aid
for the Infirm and Elderly Blind

• outdoor navigation PLUS physically support
• Labmate mobile base
• Joystick
• Polaroid sonar sensors

• Infrared proximity sensors
• command bar with Braile code key
• tone and voice feedback

27
HITOMI Yamanishi University (Japan)

• hitomi pupil
• outdoor navigation PLUS physically support
• vision system
• sonar system

• DGPS and digital map
• voice MMI
• command bar with Braille code key.