Español

The E-Puck2 ARGoS Plug-in

E-Puck2 Robot

We present a new plug-in for the ARGoS swarm robotic simulator to implement the E-Puck2 robot model, including its graphical representation, sensors and actuators. We have based our development on the former E-Puck robot model (version 1) by upgrading the existing sensors (proximity, light, ground, camera, and battery) and adding new ones (time of flight and simulated encoders) implemented from scratch. We have adapted the values produced by the proximity, light and ground sensors, including the E-Puck2’s onboard camera according to its resolution, and proposed four new discharge models for the battery. We have evaluated this new plug-in in terms of accuracy and efficiency through comparisons with real robots and extensive simulations. In all our experiments the proposed plug-in has worked well showing high levels of accuracy. The observed increment of execution times when using the studied sensors varies according to the number of robots and types of sensors included in the simulation, ranging from a negligible impact to 53% longer simulations in the most demanding cases.

More about this in our Open Access article: https://doi.org/10.1016/j.robot.2023.104412

Plug-in available at: https://gitlab.com/uniluxembourg/snt/pcog/adars/e-puck2

Robot Model

E-Puck2 simulation model: Robot’s LEDs
E-Puck2 simulation model: Robot’s LEDs

Available Sensors and Actuators

E-Puck2 ARGoS plugin: sensors and actuators
E-Puck2 ARGoS plugin: sensors and actuators

ARGoS Simulation

ARGoS simulation using the E-Puck2 plug-in
ARGoS simulation using the E-Puck2 plug-in

Publications:

  • Daniel H. Stolfi and Grégoire Danoy. Design and analysis of an E-Puck2 robot plug-in for the ARGoS simulator. In: Robotics and Autonomous Systems, vol. 164, pp. 104412, 2023.

    doi> 10.1016/j.robot.2023.104412 | [BibTex] | [Files]

This work is supported by the Luxembourg National Research Fund (FNR) - ADARS Project, ref. C20/IS/14762457. The experiments presented in this paper were carried out using the SwarmLab facility of the FSTM/DCS.

Creative Commons License