swengeler / autonomous-quadcopter-construction

This repository contains code for my Bachelor Thesis titled "Autonomous Assembly of Structures using a Quadcopter".

Autonomous construction of block structures with a swarm of quadcopters

This repository contains part of the code for my Bachelor Thesis project at the Department of Data Science and Knowledge Engineering at Maastricht University. The code implements a high-level simulation for autonomous construction using quadcopters. The res folder contains files with representations for a large number of different structures that can be built using the simulator.

Installation

The code requires Python 3. All dependencies can be installed from the included requirements.txt file. Note that the code contains some absolute file paths referring to my personal computer. Therefore the experiment utility (experiments.py) should be started with the correct command line flags for loading and saving data.

Usage

For launching a simple GUI to start a single run of simulated construction, main.py can be called from the command line or in an IDE.

The file experiments.py launches a command line interface for configuring and running experiments. The following is the help message describing its correct usage.

usage: experiments.py [-h] [-l LOAD_DIRECTORY] [-s SAVE_DIRECTORY]
                      [--skip-existing]
                      map_name

Run and save results of simulated construction with quadcopters.

positional arguments:
  map_name              Name of the map/structure to build.

optional arguments:
  -h, --help            show this help message and exit
  -l LOAD_DIRECTORY, --load-directory LOAD_DIRECTORY
                        Either a number (0-2) specifying one of three pre-
                        defined paths to load maps from or an absolute file
                        path.
  -s SAVE_DIRECTORY, --save-directory SAVE_DIRECTORY
                        Either a number (0, 1) specifying one of two pre-
                        defined paths to save results to or an absolute file
                        path.
  --skip-existing       Specifies whether existing files should be overwritten
                        if experiments are repeated. The default is to
                        overwrite them.

The functions in analysis.py were used for analysis of the data gathered from the experiments. It is currently undocumented and very messy.