SWARMS: Multi-Agent Control Simulation Platform

Multi-Agent Control Simulation Platform: Creating a customizable drone swarm control and simulation platform utilizing the cloud.

Advisor:

Description:

The primary motivation for SWARMS evolved from the need for a robust simulation platform that goes beyond standard MATLAB simulation tools. To this end, open source software will be utilized to construct an algorithmic design and testing portal that utilizes many aspects of autonomous controls, machine learning and data science. Our team will develop a web-based user portal, cloud-based simulation engine and a set of standard algorithm components for autonomous control of drone swarms. Opportunities to design, build and test machine learning models will also be provided through the creation of obstacle avoidance systems utilizing LiDAR and computer vision. An additional opportunity may also be available to simulate the Purdue Aerial Robotics autonomous flight systems team in development of their autonomous drone.

Relevant Technologies:

  • Machine Learning
  • Autonomous Controls
  • Cloud-based simulation development

Prerequisites:

  • A basic understanding of Python/C++/Java (Python & object orientated programming is a plus).
  • Familiarity with Git, software versioning tools, cloud-hosting services such as AWS, etc.
  • Perseverance, determination and team-orientated attitude.
  • Curiosity about controls, multi-agent algorithms, autonomy, machine learning and data science. Data Science principles and tooling (Pandas, Numpy, Matplotlib)
  • Web-based programming or cloud infrastructure.
  • Video game design in Unreal Engine or Unity would be an asset.
  • Timeliness, dedication and honesty.

Meeting Times:

  • Spring 2022: Mondays 11:30 am-12:20 pm, EE 013
  • Fall 2022: Mondays 11:30 am-12:20 pm, EE 013