Loading...

Design of a Distributed Controller for Stabilizing the Locomotion of Seven-Link Underactuated Planar Biped Robot with Training Ability

Kakaei, Mohammad Mehdi | 2017

566 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 50162 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Salarieh, Hassan
  7. Abstract:
  8. In this work inspiring from the nature a control method is proposed for a stable rhythmic walking in a seven-link underactuated biped robot. Stable walking is a very important issue in biped robots and proposing a dynamically stable pattern of motion with the capability of acceleration and learning is our main purpose. It is tried that the presented method make the robot have a human like motion. This method controls dynamically the hybrid model of robot’s movement and stabilizes it by converging the time-invariant constraints considered to make this movement. Moreover, in addition to providing a suitable gait for the bipod robot, a robust control method is designed to improve the ability of control’s action. This method can stabilize the motion cycle against disturbances and uncertainties in the components of the robot’s model. Besides, a control method with a hierarchical structure is designed. In this structure, a high level controller, that also called the central controller, is placed at the outer layer. The main task of this controller is stabilizing the robot movement and developing periodic trajectories while guaranteeing total stability for the entire robot dynamics. Additionally, there is a system that can learn from the produced trajectories. After training, the control system is switched to the low level controllers, which are placed in the inner layer of structure. Local controllers are in this layer and have a distributed form, which interconnect with each other as a network. By receiving local information regarding adjacent links, they individually track the desired paths for each joint. If the distributed low level controllers’ performance in total stability of the robot’s movement is not desirable, the high level controller takes its place. This control architecture has several advantages such as online learning and correction according to changes in locomotion cycles and provides a new vision in controlling humanoid robots inspired by the nature
  9. Keywords:
  10. Bipedal Robot ; Locomotion ; Control ; Stability ; Underactuated Biped Robot ; Rhythmic Movement ; Learning Ability

 Digital Object List

 Bookmark

No TOC