Presentation Title

AeroCube

Faculty Mentor

Praveen Shankar, Oscar Morales Ponce

Start Date

18-11-2017 10:00 AM

End Date

18-11-2017 11:00 AM

Location

BSC-Ursa Minor 98

Session

Poster 1

Type of Presentation

Poster

Subject Area

engineering_computer_science

Abstract

PDF: https://goo.gl/s54TvU

Collaborative autonomous systems are composed of small, interconnected components that solve complex problems by distributing work load and considering multiple perspectives. While difficult to implement, these kinds of systems are both dynamic and scalable. They have extraordinary potential in space, whose vast environment can easily overwhelm traditional systems. This is the motivation behind our research.

The goal of the AeroCube project was to create a system of robots that move into a given formation, then find and track a specified marker. From the start, it was clear that localization, communication, computer vision, path planning, and motion control would be essential. To experiment with multiple implementations for each of these components, we created a software architecture that allows for hardware and algorithm interchangeability. Currently, algorithms in use include GPS-like localization, UDP broadcasting, Aruco marker detection, and Hopcroft–Karp matching.

The high-level behavior of the system was implemented through a state machine representing three phases: formation, scan, and track. In the formation phase, robots calculate and move to their respective positions in a given formation. Then, they enter the scan phase and iteratively rotate in one direction until the specified marker is detected. Upon detection, each robot independently enters the track phase and rotates to center the marker in its frame of reference. If the marker is lost, the scan phase is re-entered.

Moving forward, we will explore more collaborative decision making, robust object detection, and more. The AeroCube project lays the groundwork for future research in collaborative autonomous systems and their applications.

This document is currently not available here.

Share

COinS
 
Nov 18th, 10:00 AM Nov 18th, 11:00 AM

AeroCube

BSC-Ursa Minor 98

PDF: https://goo.gl/s54TvU

Collaborative autonomous systems are composed of small, interconnected components that solve complex problems by distributing work load and considering multiple perspectives. While difficult to implement, these kinds of systems are both dynamic and scalable. They have extraordinary potential in space, whose vast environment can easily overwhelm traditional systems. This is the motivation behind our research.

The goal of the AeroCube project was to create a system of robots that move into a given formation, then find and track a specified marker. From the start, it was clear that localization, communication, computer vision, path planning, and motion control would be essential. To experiment with multiple implementations for each of these components, we created a software architecture that allows for hardware and algorithm interchangeability. Currently, algorithms in use include GPS-like localization, UDP broadcasting, Aruco marker detection, and Hopcroft–Karp matching.

The high-level behavior of the system was implemented through a state machine representing three phases: formation, scan, and track. In the formation phase, robots calculate and move to their respective positions in a given formation. Then, they enter the scan phase and iteratively rotate in one direction until the specified marker is detected. Upon detection, each robot independently enters the track phase and rotates to center the marker in its frame of reference. If the marker is lost, the scan phase is re-entered.

Moving forward, we will explore more collaborative decision making, robust object detection, and more. The AeroCube project lays the groundwork for future research in collaborative autonomous systems and their applications.