ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-06292016-182801


Tipo di tesi
Tesi di laurea magistrale
Autore
TONDO, ALESSANDRO
URN
etd-06292016-182801
Titolo
Humanoid Robot Modeling and Implementation of a Walking Control in Simulated Environment
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
INGEGNERIA ROBOTICA E DELL'AUTOMAZIONE
Relatori
relatore Prof. Caiti, Andrea
correlatore Prof.ssa Laschi, Cecilia
correlatore Dott. Falotico, Egidio
Parole chiave
  • ros
  • humanoid robot locomotion
  • gazebo simulator
  • closed-form inverse kinematics
  • biped dynamic walking
  • task-priority inverse kinematics
  • zero-moment point
Data inizio appello
21/07/2016
Consultabilità
Non consultabile
Data di rilascio
21/07/2086
Riassunto
Even if it has been studied since the early '70s the locomotion task in the humanoid robotics field is still an actual and challenging research topic. The complexity of a human-like dynamic walk process involves almost every aspects of robotics: from the mechanical project and development of the most suitable robot structure, to the control capabilities including stereo vision aid, footstep planning, stabilization and a huge variety of robust control techniques.
Due to this extent I have chosen to focus the analysis only on the crucial low level control of the locomotion task, assuming that all the desired operative space trajectories --- i.e. the walking gait --- are provided by an ideal planner. Nonetheless the whole development of this thesis work is entirely in simulation environment.

The SABIAN (Sant'Anna BIped humANoid) robot, developed in conjunction with the Waseda University of Tokyo, is the humanoid robot of interest, which is modeled to mimic the real robot behavior in the Gazebo simulator and which is used in this dissertation for the locomotion task.
In particular, I present two distinct approaches for the online computation of the inverse kinematics of a highly redundant system --- i.e. the humanoid robot --- and one simple control algorithm which exploit them to perform a human-like dynamic walk with stretched knees. Simulation results are discussed in the last part of this thesis.
File