Loading...

Dynamics Modelingand Decentralized Control of Scalable Multi Agent Holonomic Systems

Sadeghian, Mosayeb | 2014

517 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 46235 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Sayyaadi, Hassan
  7. Abstract:
  8. Today, a new approach in the use of robots and the division of task between them to accomplish a specific mission or doing group exercises, has emerged. In fact, the main idea is that the multi agent robotic systems, soon is the most economically advantageous and will take variety of tasks that require distribution understanding of the environment (eg, search and rescue, surveillance, environmental monitoring and military operations). In recent years, the autonomous wheeled mobile robots, OMR, have special significance. This type of robots for independent control of translational and rotational motion, have the holonomic behavior dynamic that this means unlike other nonholonomic robots, be able to move instantaneously and without delay in any direction that expedite the operations.
    The purpose of this project is the scalable and decentralized control of a group of robots to collision avoidance with each other and formation control of complex in a special geometrical form and Swarm tracking. Using lagrangian method, the nonlinear equations of each agent is extracted. In the applied control algorithm, using potential field method, the proper control inputs for collecting the agents around the desired geomtry by maintaining the certain equilibrium distance (formation control) is designed. In addition, the weighted graph concepts is used to control formation of a group of agents in the presence of environmental obstacles. Then, sliding mode control is used for applying a lower level controller for each robot. To solve the problem of positioning of agents in the local minimums, which leads to the of undesirable formation, virtual relabeling algorithms is used. Finally, the simulation results show the efficiency of the algorithm.
  9. Keywords:
  10. Multiagent System ; Formation Control ; Potential Field Method ; Weighted graphs ; Decentralized Controller ; Virtual Relabeling

 Digital Object List

 Bookmark

No TOC