Performance Indicators of Humanoid Posture Control and Balance Inspired by Human Experiments

EasyChair Preprint 4268

Abstract

Posture control and maintaining balance are fundamental elements of humanoid robot control and have a significant impact for the performance of robots. The evaluation of robotic performance, at the state of the art, is mostly evaluated at goal level, e.g. with robot competitions. While falling is a typical reason beyond the failure of the humanoid operation, the failure itself does not provide many details about the nature of the underlying problem that can be used to improve the control. In order to provide a more specific analysis of posture control and balance, this contribution presents a set of performance indicators, i.e. indexes that can be used to compare the performance of robots with the human control systems. The inspiration for the proposed tests and indicators comes from human experiments and particular emphasis is placed on human-robot comparison.

Keyphrases: Benchmarking, Humanoids, balance, human-like, posture control

@booklet{EasyChair:4268,
  author    = {Vittorio Lippi and Thomas Mergner and Christoph Maurer and Thomas Seel},
  title     = {Performance Indicators of Humanoid Posture Control and Balance Inspired by Human Experiments},
  howpublished = {EasyChair Preprint 4268},
  year      = {EasyChair, 2020}}