A SwarmAI Testbed for Workforce Development and Collaborative, Interdisciplinary Research

Martha Cervantes; Raphael Norman-Tenazas; Adam Goertz; Erik C. Johnson; William Roberts Gray-Roncal

Download Paper | Permalink

Conference: 2023 ASEE Annual Conference & Exposition
Location: Baltimore , Maryland
Publication Date: June 25, 2023
Start Date: June 25, 2023
End Date: June 28, 2023
Conference Session: COED Modulus Topics
Tagged Division: Computers in Education Division (COED)
Page Count: 8
DOI: 10.18260/1-2--42515
Permanent URL: https://strategy.asee.org/42515
Download Count: 129

Request a correction

Paper Authors

biography

Martha Cervantes Johns Hopkins University

visit author page

Martha Cervantes is a Mechanical Engineer at the Johns Hopkins University Applied Physics Laboratory where she works in mechanical design and integration of robotic systems. Additionally, Martha is the project manger of the CIRCUIT Program at JHU/APL, which connects and mentors students from trailblazing backgrounds to STEM careers through science and engineering projects. Martha received her B.S. in Mechanical Engineering from Johns Hopkins University, and she is currently pursuing a M.S. in Mechanical Engineering at the Johns Hopkins University.

visit author page

author page

Will Gray-Roncal, PhD, is a Group Supervisor and Senior Research Engineer at the Johns Hopkins University Applied Physics Laboratory. His research interests include Learning Science, advancing equity and inclusion in STEM, and Connectomics.

visit author page

Download Paper | Permalink

Abstract

Educating the next generation of AI researchers requires methods which teach the software tools, theoretical concepts, and domain knowledge specific to the field. To help develop these key skills, we focus particularly on the area of Swarm AI, which, in general, covers the autonomous operation of a large number of agents in a single environment. Applications of this field are numerous, including autonomous navigation, defense, robotics, logistics, and search/rescue. Despite the potential for impact in key problem domains and interdisciplinary nature, Swarm AI platforms are note generally used to develop competency in AI education. While some courses and tracks target AI in general, there are few tools to help engage learners in the specific area of Swarm AI. This is partly due to needing strong skills in the intersection of reinforcement learning, software development, and robotics, as well as a framework computing capability on which to test and evaluate. We propose Battle Optimized Laser Tag (BOLT), a Unity-built simulator that enables learners to develop Swarm AI algorithms in the context of a laser tag game between small mobile robots. The simulator is built to be familiar to a video game environment to maximize engagement, while the included documentation for the environment is designed as a gentle introduction into the topic of controlling swarms. Easily modified configurations enable educators to specify tasks and develop curricula to further challenge learners. Furthermore, the simulator is OS-agnostic, simple to install and uses Python to interface with the agents. This allows learners to implement their solutions on their own computers in a programming language common to other AI courses. Finally, extensions of the platform, leveraging the Python interface, can be deployed to physical robots or be used as a testbed for other AI domains beyond robotics including Human-Machine Teaming and Cybersecurity.

Citation
Format

Cervantes, M., & Norman-Tenazas, R., & Goertz, A., & Johnson, E. C., & Gray-Roncal, W. R. (2023, June), A SwarmAI Testbed for Workforce Development and Collaborative, Interdisciplinary Research Paper presented at 2023 ASEE Annual Conference & Exposition, Baltimore , Maryland. 10.18260/1-2--42515