ETD system

Electronic theses and dissertations repository


Tesi etd-09062019-142613

Thesis type
Tesi di laurea magistrale
Mechatronic Modelling, Control and Test of Bipedal Robot Walking on Compliant Feet
Corso di studi
relatore Prof. Bicchi, Antonio
relatore Prof.ssa Venture, Gentiane
relatore Catalano, Manuel Giuseppe
relatore Grioli, Giorgio
Parole chiave
  • Soft Robotics
  • Humanoid Robot
  • Robotic Foot
Data inizio appello
secretata d'ufficio
Data di rilascio
Riassunto analitico
The majority of humanoid robots existing in literature adopt flat feet, a choice that can limit their performance when maneuvering over uneven terrains. Softfoot is a soft robotic foot, inspired to a human foot, that allows greater adaptability to the ground. In the first part of this thesis, carried out at GVLab of Tokyo University of Agriculture and Technology (TUAT), we analyzed the deformation of the plantar fascia of the Softfoot robotic foot, comparing it with the behavior of the human foot. Through Motion Capture it was possible to reconstruct the longitudinal deformation of both types of foot (along the walking direction), and the transverse deformation. We evaluated the adaptability of the robotic foot on obstacles and observed the comparison with the human foot. In the second part of the thesis, carried out at National Institute of Advanced Industrial Science and Technology (AIST), Japan, a soft robotic foot designed to adapt to the ground was proposed to overcome part of these limitations. This thesis presents the results of testing two such feet on the humanoid robot HRP-4, and compares them to what obtained with the original flat feet of the robot. After describing the soft foot and how it has been adapted to the robot, the biped is tested while balancing, stepping and walking. Tests are carried out on flat ground and on obstacles of different heights. For comparison purposes, the original HRP-4 controller has been used for both types of feet (except for re-evaluation of the CoM position). Analysis of the ankle pitch angle, ankle pitch torque, waist roll angle and waist pitch angle, show a substantial improvement in obstacle negotiation performance of HRP-4, when using the Softfoot, even without optimizing the controller to exploit the Softfoot features. Finally, in the last part, a static mathematical model was developed to reconstruct not only the shape of the foot, hinting at the shape of the ground beneath, but also, under precise hypotheses, the contact force distribution. The problem is theoretically formalized and analyzed in the simplified scenario derived from considering only the effects on the sagittal plane. Then, theoretical results are experimentally validated.