Gal A. Kaminka: Publications

Sorted by DateClassified by Publication TypeClassified by TopicGrouped by Student (current)Grouped by Former Students

Stable Humanoid Whole Body Motion Generation

Sharon Yalov-Handzel. Stable Humanoid Whole Body Motion Generation. Ph.D. Thesis, Bar Ilan University, 2016.

Download

[PDF]1.7MB  

Abstract

The stability of humanoids is fragile and depends on the robot dynamics which are difficult to plan in advance. Keeping humanoids stable along their non-repetitive motion sequences is a great challenge in robotics, and it is usually resolved by using feedback control to consider the dynamic impacts. Off-line planning of whole body stable motion cannot be resolved by motion generators based on feedback control. We propose an algorithm to plan stable whole body postures, based on an IKP (Inverse Kinematics Problem) solver. The IKP is extended to solve the kinematic equations under any condition that can be represented in geometrical form. Such a condition is the robot stability. This numerical iterative algorithm can be applied to any robot structure. In addition, the algorithm can find a kinematic solution that obey a certain optimization criteria, but is not complete. In order to cope with the algorithm's incompleteness, we propose some improvements that increase the convergence probability of the solution. The algorithm was analyzed and simulated and the core contribution of this work is that it gives a general method for off-line planning of complex postures applied to high DOF (degrees of freedom) robots that should obey some condition. Moreover, the improvements demonstrate that there is a tradeoff between the completeness and the generality of the solver.

Additional Information

BibTeX

@PhdThesis{yalov-phd, 
author = {Sharon Yalov-Handzel}, 
title = {Stable Humanoid Whole Body Motion Generation}, 
school = {{B}ar {I}lan {U}niversity}, 
year = {2016}, 
  wwwnote = {}, 
OPTannote = {} ,
  abstract = {The stability of humanoids is fragile
  and depends on the robot dynamics which are difficult
  to plan in advance. Keeping humanoids stable along
  their non-repetitive motion sequences is a great challenge
  in robotics, and it is usually resolved by using
  feedback control to consider the dynamic impacts. Off-line
  planning of whole body stable motion cannot be resolved
  by motion generators based on feedback control. We 
  propose an algorithm to plan stable whole body postures,
  based on an IKP (Inverse Kinematics Problem) solver.
  The IKP is extended to solve the kinematic
  equations under any condition that can be represented
  in geometrical form. Such a condition is the robot
  stability. This numerical iterative algorithm can be
  applied to any robot structure. In addition, the algorithm
  can find a kinematic solution that obey a certain
  optimization criteria, but is not complete. In order to
  cope with the algorithm's incompleteness, we propose
  some improvements that increase the convergence
  probability of the solution. The algorithm was analyzed
  and simulated and the core contribution of this work
  is that it gives a general method for off-line planning
  of complex postures applied to high DOF (degrees of freedom) robots
  that should obey some condition. Moreover, the improvements
  demonstrate that there is a tradeoff between the completeness
  and the generality of the solver.
 },
}

Generated by bib2html.pl (written by Patrick Riley ) on Fri Aug 30, 2024 17:29:51