CRIM Logo
Center for Robotics and Intelligent Machines

Research

Always on the cutting edge of advanced robotics and intelligent machines research, CRIM members are constantly creating, cultivating, and completing projects that delve the depths of new and exciting ideas and technology. CRIM's inspired insights into the near-future of robotics and intelligent machines are presented here for public consumption.
CRIM C-5 Galaxy Wall-Climbing Rivet Inspector

CRIM C-5 Galaxy Wall-Climbing Rivet Inspector

Projects Toolkit Icon

Shape Formation of Robot Swarms

Project Researcher(s): 
Nikhil Deshpande

The idea of multiple robots displaying group behavior has been explored for many years. Group behavior in robots is a desired feature as, for the execution of a particular complex task, it allows for a reduction in the complexity of an individual robot while still exhibiting the same level of group intelligence, required for completing the task itself, e.g., in region exploration task in a search and rescue operation. Similarly, for inspecting rivets on an aircraft fuselage, the shape of the group must be maintained.

Figure 1 Robot communication nodeA coordinated group of robots can assist rescuers when going into inaccessible areas, e.g., where the terrain is unknown. Having multiple robots will allow the group to spread over a larger region, thus, simultaneously sensing different parts of the region. The group behavior in robots will allow them to move efficiently over the terrain while maintaining proximity and information exchange capabilities. Different approaches exist for the implementation of these group patterns including: (1) biological, and (2) physical. Another artifact of the group behavior is the requirement of some form of communication amongst the robots to maintain group integrity. The recent development in low cost, low power wireless communication technology makes it a prime technology for adopting for sensing and controlling robot swarms. Different features in wireless technologies such as signal intensity and signal quality can now be used to dictate the behavior of the robot group.

The ability of the robot swarm to adapt to the terrain dynamically while still maintaining group integrity is the target of this research into robotic group behavior. In a task such as moving an object through a maze, or a Figure 2 An autonomous mobile sensor platformmore natural setting of search & rescue in an unknown environment, a robot swarm shall acquire an optimum shape suitable for Figure 3 Wireless signal intensity calculation tests performed with directional antennasexecuting the task. While proceeding to execute it, the group needs to sense the obstacles in the environment, adjust the individual positions of the robots to give a different but locally optimal shape to the swarm, then proceed through that location and reconvene into the original optimum shape.

The platforms use wireless technology that includes the MSP430 microcontroller technology with Chipcon radio transceivers for the wireless communication. The robotic platform for the experiments initially involves a simple wheeled vehicle with the attached wireless infrastructure for the control and communication logic.