Overweight Object Transportation with a Set of Collaborative Robots

Abstract

We study the box pushing problem with a set of robots. Robots and relocated objects are placed on opposite sides in such a way that robots have a direct view of the objects. We consider a rather weak model where robots do not have access to a localization system and are disoriented. Further, robots do not know the weight of the boxes nor of their own pushing capabilities. Robots do have access to a wireless device to communicate with other robots. The objective is to design a distributed algorithm that lets the robots self-coordinate to move the objects. We propose a synchronous algorithm that allows the robots to self-coordinate to relocate the box. We use Timed Input/Output Automata to describe the algorithms and show that the algorithm correctly completes the task. Then we extend the algorithm to deal with obstacles and robot failures. We implement the algorithm in SyRof (a testbed built at California State University Long Beach.) The testbed consists of four robots equipped with omnidirectional wheels to simulate drones and an autopilot that provides a synchronous system. The resulting implementation allows the robots to complete the task successfully.