Efficient Bipedal Robots Based on PassiveDynamic Walkers Review - PowerPoint PPT Presentation

1 / 18
About This Presentation
Title:

Efficient Bipedal Robots Based on PassiveDynamic Walkers Review

Description:

They are simple mechanical devices, composed of solid parts connected by joints, ... reduced in two steps to 6 atm through locally developed miniature pneumatics ... – PowerPoint PPT presentation

Number of Views:117
Avg rating:3.0/5.0
Slides: 19
Provided by: S726
Category:

less

Transcript and Presenter's Notes

Title: Efficient Bipedal Robots Based on PassiveDynamic Walkers Review


1
  • Efficient Bipedal Robots Based onPassive-Dynamic
    Walkers Review
  • Jianhua (Sam) Xie
  • 05172055

2
Overview
  • Passive-dynamic walkers
  • They are simple mechanical devices, composed of
    solid parts connected by joints, that walk stably
    down a slope.
  • They have no motors or controllers and can have
    humanlike motion

3
Overview-Different approach
  • Most researcher study human walking by measuring
    joint angle and reaction force
  • passive-dynamics model compares to human in terms
    of morphology, gait appearance, energy use, and
    control
  • The robots do not control any joint at any time

4
Overview
  • The article present three robots based on
    passive-dynamics, with small active power sources
    substituted for gravity, which can walk on level
    ground.
  • These robots use less control and less energy
    walk more naturally

5
Passive designs
  • The Wilson Walkie
  • MITs improved version
  • Cornell copy of McGeers capstone design
  • The Conell passive biped with arms

6
Three Robots Developed
  • Three robots using passive-dynamic machines were
    built in at three institutions
  • The Cornell biped (A)
  • The Delft biped (B)
  • MIT learning biped (C)

7
The Cornell biped
  • Machine weight 12.7kg and has 5 internal degrees
    of freedom
  • The robot has only one capability, walking
    forward
  • Low energy use (primary goal)
  • A proof-of-concept prototype

8
The Cornell biped
  • The Cornell biped is specifically designed for
    minimal energy use.
  • The primary energy losses for biped walking at a
    constant speed is due to dissipation
  • Cornell design demonstrates that its possible to
    avoid this negative actuator work

9
The Cornell biped - how to walk
  • One hip, two knees, and two ankles
  • Ankle push-off restores energy lost
  • State switching
  • Left Leg Swing
  • Right Ankle Push-off
  • Right Toe Return
  • Right Leg Swing

10
The Delft biped
  • Machine weight 8kg and has 5 internal degrees of
    freedom
  • Air-actuated artificial muscle (McKibben muscles)
  • A proof-of-concept prototype

11
The Delft biped
  • The muscles are fed with CO2 from a 58 atm
    cannister, pressure-reduced in two steps to 6 atm
    through locally developed miniature pneumatics

12
The Delft biped
  • The success of the robot at balancing using
    ankles
  • swift swing-leg motion not only increased
    fore-aft stability but also increased lateral
    stability

13
MIT learning biped
  • Machine weight 2.57kg and has 6 internal degrees
    of freedom
  • The robot is designed to test the utility of
    motor learning on a passive-dynamic mechanical
    design.

14
MIT learning biped
  • The goal of the learning is to find a control
    policy that stabilizes the robots trajectory on
    level terrain using the passive ramp-walking
    trajectory as the target.

15
MIT learning biped
  • Measurements from previous and current step to
    efficiently estimate the performance gradient on
    the real robot
  • despite sensor noise
  • imperfect actuators and uncertainty in the
    environment

16
MIT learning biped
  • The MIT biped shows that the efficiency of motor
    learning can be strongly influenced by the
    mechanical design of the walking system, both in
    robots and possibly in humans.

17
Three robots mimic humans
  • Each of robots here has some design features that
    are intended to mimic humans
  • Cornell and Delft use anthropomorphic geometry
    and mass distributions in their legs and
    demonstrate ankle push-off and powered leg
    swinging
  • MIT biped uses a learning rule that is
    biologically plausible at the neural level

18
Conclusion
  • The conclusion that natural dynamics may largely
    govern locomotion patterns was already suggested
    by passive-dynamic machines.
Write a Comment
User Comments (0)
About PowerShow.com