Cooperation in a Heterogeneous Robot Swarm through Spatially Targeted Communication

Nithin MathewsAnders Lyhne ChristensenRehan O'Grady and Marco Dorigo

Abstract


We consider a heterogeneous swarm robotic system composed of wheeled and aerial robots called foot-bots and eye-bots, respectively. The foot-bots can manipulate objects on the ground. They are also able to physically connect to one another autonomously and thus form larger, composite robotic entities. Eye-bots have a privileged overview of the environment since they can fly and attach to metal ceilings. In this paper, we show how the heterogeneous swarm can benefit from cooperation. By using so called spatially targeted communication, the eye-bot is able to communicate with selected groups of foot-bots and instruct them on how to overcome obstacles in their path by forming morphologies appropriate to the object encountered. We conduct experiments in simulation to quantify separately the benefits of cooperation and of spatially targeted communication.


Non-cooperative control (NCC)


The foot-bots are provided with a SWARMORPH-script that enables formation a line of 4 foot-bots when a gap is encountered regardless of the width of the gap. The foot-bots do not communicate nor do they attempt to cooperate with an eye-bot. They initially move towards the light source until one of the foot-bots detects the gap. The foot-bot retreats approximately 40cm from the gap and becomes the seed by inviting other foot-bots to connect to its rear. Other foot-bots that are able to see the connection slot stop moving towards the light source and attempt to connect to the morphology. Once the line of 4 foot-bots is formed, the foot-bots move towards the light source to cross the gap.


Gap width 5cm

Gap width 10cm

Gap width 15cm

Gap width 25cm


Cooperative control with spatially targeted communication (CC-STC)


The foot-bots do not have a priori knowledge about the task or the target morphology. The foot-bots initially move towards the light while avoiding the gap until the eye-bot illuminates its red LEDs. Spatially targeted communication link is then established with optimally situated foot-bots as described in the paper. The eye-bot selects a foot-bot that is approximately 40cm away from the gap to be the seed leaving enough room for the free foot-bots to maneuver around the morphology during the self-assembly process. Then, a certain number of immediate neighbors of the seed robot are selected to form the target morphology. The selected foot-bots receive the SWARMORPH-script from the eye-bot and follow the instructions in the script to self-assemble into the target morphology. Once the morphology is formed, the foot-bots move towards the light to cross the gap. The non-selected foot-bots move away from the light.


Gap width 5cm

Gap width 10cm

Gap width 15cm

Gap width 25cm


Cooperative control with random group selection (CC-RGS)


The foot-bots do not have a priori knowledge about the task or the target morphology. The foot-bots initially move towards the light until the eye-bot illuminates its red LEDs. The eye-bot then randomly selects a seed foot-bot (i.e., without considering its location in the environment with respect to the gap) and further foot-bots (i.e., without considering their relative distance to the seed foot-bot) required to form the target morphology. The selected foot-bots receive the SWARMORPH-script from the eye-bot and follow the instructions in the script to self-assemble into the target morphology. Once the morphology is formed, the foot-bots move towards the light to cross the gap. The non-selected foot-bots move away from the light.


Gap width 5cm

Gap width 10cm

Gap width 15cm

Gap width 25cm