Abstract

Swarm of biological species exhibits a variety of simple individual characteristics which collectively give rise to complex behaviors such as group formation and transportation. A Decentralized control-based algorithm is presented wherein using each robot in the swarm uses local control rules to exhibit different behavior that results in collective transportation. On the basis of this implicit coordination through applying force towards the same direction the robot swarm successfully manipulates the object to the goal location. Collective/emerging behavior is achieved through local control of the robots as hown in the fig 1.

                                                                               Fig 1: Basic Working Principle

Assumptions and Constraints

• The object (load) is assumed to be rigid and have an evenly distributed mass.
  
• Robots do not know the position of each other in the environment.
  
• The object mass is known, whereas the object geometry is not known.
  
• All the robots are applying force in the same direction, so the total external force will also have the same direction.
  

Conditions that need to be satisfied -

• Locomotion- This can be achieved with differential wheels.
  
• Sense the orientation of the goal- This can be done by using cameras, photodiodes.
  
• Attach to the object while not losing freedom in movement- Standard grippers mounted on omnidirectional base.

Problem Formulation and Model

                                                                                                Fig 2: Model