Establishing Spatially Targeted Communication in a Heterogeneous Robot Swarm

Nithin MathewsAnders ChristensenEliseo FerranteRehan O'Grady and Marco Dorigo

Abstract


We consider a heterogeneous swarm of aerial robots and wheeled robots. We present a system that enables spatially targeted communication without relying on any form of global information. Our system allows an aerial robot to establish a dedicated communication link with individual wheeled robots or with selected groups of wheeled robots based on their position in the environment. We show how a spatially targeted one-to-one communication link can be established using a simple LED and camera based communication modality. We provide a probabilistic model of our approach to derive an upper bound on the average time required for establishing communication. We show in simulation- based experiments that our approach scales well. Furthermore, we show how our approach can be extended to establish a spatially targeted one-to-many communication link between an aerial robot and a specific number of co-located wheeled robots. The heterogeneous swarm robotic hardware is currently under development. We therefore demonstrate the proposed approach on an existing multirobot system consisting of only wheeled robots by letting one of the wheeled robots assume the role of the aerial robot.


One-to-one Communication



Simulation 1: 10 wheeled robots, 2 colors in the selection process, 4 iterations
This video footage shows the simulation-based experimental setup discussed in the paper. This particular simulation starts with an aerial robot placed in the center of a closed, obstacle-free arena (2m x 2m) at a height of 2m. Ten wheeled robots are randomly placed within the visual range of this aerial robot. The aerial robot is able to perceive all wheeled robots within the arena and vice versa. A total of 3 colors (red, green and blue) are available to both the aerial robot and the wheeled robots. The color red is used to initiate and confirm the termination of the process. The colors green and blue are used in the selection process. For the sake of better visibility, the colors displayed on the LEDs are reinforced on the ground.

In the following, we provide further video footage of the simulation-based experiments carried out to test the scalability of our approach. In the case of six colors in the selection process, the colors green, blue, yellow, cyan, violet and white are used.

Video Wheeled robots No. of colors in the selection process Iterations
Simulation 2 10 6 3
Simulation 3 20 2 6
Simulation 4 20 6 3
Simulation 5 40 2 8
Simulation 6 40 6 5
Simulation 7 60 2 8
Simulation 8 60 6 3
Simulation 9 80 2 12
Simulation 10 80 6 4





Experiment 1: 4 wheeled robots, 2 colors in the selection process, 4 iterations
This is the video footage of the real robot experiment described in the paper. The s-bot on the bottom assumes the role of the aerial robot and establishes a one-to-one communication link with the s-bot on the top-left. A total of 3 colors (red, green and blue) are used in the control program. The color red is used to initiate and confirm the termination of the process. The colors green and blue are used in the selection process. The selection process is iterated four times before a one-to-one communication link is successfully established.

In the following, we provide further video footage of similar proof-of-concept experiments. The number of wheeled robots is kept constant at 4 and the number of colors in the selection process at constant 2. Also listed is the number of iterations of the selection process.

Video Wheeled robots No. of colors in the selection process Iterations
Experiment 2 4 2 2
Experiment 3 4 2 2
Experiment 4 4 2 3
Experiment 5 4 2 4
Experiment 6 4 2 3
Experiment 7 4 2 3
Experiment 8 4 2 4
Experiment 9 4 2 2
Experiment 10 4 2 3


One-to-many Communication



Simulation 1: Group size 10, approximate growth, 2 iterations, 14 excess robots
This video footage shows the simulation-based experiments explained in the paper. The simulation starts with an aerial robot placed in the center of a closed, obstacle-free arena (2m x 2m) at a height of 2m. Eighty wheeled robots are randomly placed within the visual range of this aerial robot. The aerial robot is able to perceive all wheeled robots within the arena and vice versa. Furthermore, the wheeled robots are now able to perceive neighboring wheeled robots within a radius of 1m. Six colors are used in the selection process while establishing the one-to-one communication link. For the sake of better visibility, the colors displayed on the LEDs are reinforced on the ground. In a first phase, the footage shows how a one-to-one communication link is established to a particular wheeled robot. In a second phase, this one-to-one communication link is extended (or grown) to a one-to-many communication link to include a minimum of 10 robots. This is achieved within 2 iterations including an excess of 14 robots.

In the following, we provide further video footage of the two types of simulation-based experiments we ran on the establishment of one-to-many communication: (1) an exact group size is required, and (2) the final number of robots in the group is allowed to exceed the number of robots initially desired. In the second case (approximate growth), also the number of excess robots is listed.

Video Group size Type Iterations Excess robots
Simulation 2 10 Exact 6 -
Simulation 3 20 Approximate 2 4
Simulation 4 20 Exact 4 -
Simulation 5 40 Approximate 4 12
Simulation 6 40 Exact 10 -
Simulation 7 60 Approximate 5 11
Simulation 8 60 Exact 6 -
Simulation 9 80 Approximate 6 0
Simulation 10 80 Exact 6 -





Experiment 1: Group size 2, exact growth, 1 iteration
This is the video footage of the real robot experiment described in the paper. We placed 4 s-bots in the shape of an arch around a predesignated s-bot which assumes the role of the aerial robot and seeks to grow a group of size 2. In a first phase, a one-to-one communication link is established to the rightmost s-bot. In a second phase, this communication link is expanded to become a one-to-many communication link including the second s-bot from the right.

In the following we provide further video footage of similar proof-of-concept experiments, in which the group size is varied between 2 and 3. Also listed is the number of iteratons.

Video Group size Type Iterations
Experiment 2 2 Exact 1
Experiment 3 3 Exact 2
Experiment 4 3 Exact 4