RVSS Workshop 2020

Find the Animals!

The animals have escaped Canberra zoo and run into Aruco forest! You must work together in teams and use your robots to find where they are hiding!

The zoo has hired your team along with 2 other teams to develop a robotic system to find the animals so they can be brought back safely. You must work together with the other teams to combine your search efforts, and use robotic vision technologies including SLAM, neural networks and multi-view optimisation to report the animal locations with maximum accuracy!

Workshop Overview

In this workshop, you will work in a team of 3 members to develop a robotic system to localise toy animals spread throughout an environment including numerous ArUco landmarks. On the final day of the workshop, your system will be put to the test. You will use a keyboard based control system to steer the robot, a 2D SLAM system to generate a map of the landmarks, a neural network to identify animal locations in images, and a data fusion algorithm to combine your map with the maps from two other teams and reveal the true locations of the animals.

A lot of code and systems are provided by us, but there is plenty of work to be done to make your system the very best! Since you will be working together with other teams on the final day, it is highly encouraged you collaborate and share code and ideas with each other.

When you are ready, get started by setting up your workspace.

Workshop Contents

The following links will explain how to go about completing the workshop challenge and provide hints and instructions on how to develop the code. You are encouraged to visit the links in order, although you may complete the tasks in any order you choose.

Control

Calibration

Datasets

SLAM

Neural Network


For details on how your system will be evaluated, see the Final Challenge page.