This project was developed for my bachelor thesis. Multiple algorithms for antbots are implemented in a simulator and tested for their effiecency in covering areas. Overall there are 5 Algorithms implemented:

• Node Counting
• Learning Real Time A* (LRTA*)
• Wagner Algorithm
• Batalin-Sukhatme Algorithm
• Randomwalk

The behavior of these algorithms is tested in 3 different terrains with different complexity. Furthermore the robots' behavior in swarms of different size is examined. The configuration of the simulation can be done via YAML files (see Execution).

The data is captured during runtime, saved in .dat files and can be plotted using the included python scripts.

Attention: This project was only tested under Linux Mint 17. However, it should also work on Ubuntu 14.04.

The following dependencies are required to run and compile the project:

• Stage
• yaml-cpp 0.5.1 (in lib/)
• easylogging++ 9.75 (in lib/)
• gcc 4.8 or higher

For compiling the 3 libraries the following components have to be installed via APT:

sudo apt-get install build-essential git subversion cmake pkg-config libglu1-mesa-dev libpng12-dev libjpeg8-dev libfltk1.1-dev libltdl-dev libboost-dev libboost-thread-dev

The dependencies yaml-cpp and easylogging++ can be installed using the get-deps shell script in scripts/. If the script does not work, you have to install those libraries manually.

Yaml-pp is expected to be in the lib/yaml-cpp/ subdirectory of the repository. You can find yaml-cpp here. Execute the following commands to install yaml-cpp:

cd <path-to-repo>
mkdir -p lib/yaml-cpp
cd lib/
wget "https://yaml-cpp.googlecode.com/files/yaml-cpp-0.5.1.tar.gz" -O "yaml-cpp.tar.gz"
tar xzf yaml-cpp.tar.gz -C yaml-cpp
rm -f yaml-cpp.tar.gz
cd yaml-cpp
mkdir -p bin
cd bin
cmake ..
make

The compiled libraries should be in lib/yaml-cpp/bin. Alternatively you can use build/ or lib/ as build directories.

Easylogging++ is a header-only logging framework and is expected to be in the lib/easylogging++/ subdirectory of the repository. You can find it here. Execute the following commands to install Easylogging++:

cd <path-to-repo>
mkdir -p lib/easylogging++
cd lib/
wget "https://github.com/easylogging/easyloggingpp/releases/download/v9.75/easyloggingpp_v9.75.tar.gz" -O "easylogging++.tar.gz"
tar xzf easylogging++.tar.gz -C easylogging++
rm -f easylogging++.tar.gz

The header file easylogging++.h should be in lib/easylogging++/.

Unlike the other libraries Stage will be installed globally. You can find the Stage repository here. For Stage installation the following dependencies are required:

• FLTK >=1.1
• git
• cmake
• pkg-config
• libpng
• jpeglib
• OpenGLU

Compile and install Stage (used commit f9c59ca77):

cd <any-dir>
git clone https://github.com/rtv/Stage.git
cd Stage
git checkout f9c59ca77f203049979dcf32223e5d2b26c03bf2
mkdir build
cd build
cmake ..
make
sudo make install

If you are using a 64bit linux, the Stage libraries will be installed to /usr/local/lib64 and Stage will fail to load the libraries during runtime. Also the file stage.pc will be installed to /usr/local/lib64/pkgconfig and pkgconfig won't be able to find the Stage libs for linking. Add the following lines to your ~/.bashrc to solve these problems:

# Stage
export PATH=$PATH:"/usr/local/lib64"
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:"/usr/local/lib":"/usr/local/lib64"
export STAGEPATH="/usr/local/lib":"/usr/local/lib64"
export PKG_CONFIG_PATH=$PKG_CONFIG_PATH:"/usr/local/lib64/pkgconfig"

To verify your installation, you can try to run the following scenarios. <stage-install-dir> is in most cases /usr/local/share/stage/.

# Start stage
cd <stage-install-dir>/worlds
stage simple.world

# Another Test
stage fasr.world

Because this project is compiled with C++11 , then stage.hh will give you many warnings about literals. In C++11 strings should not be concatened like this "text1"strVar"text2". Instead it should be like this "text1" strVar "text2".

To get rid of these warnings open the file /usr/local/include/Stage4.1/stage.hh and replace "m" with " m ". All warnings should be gone now.

After installing all the dependencies, compilation should be no problem. Checkout the source code and execute the following commands:

cd <path-to-repo>
mkdir build
cd build
cmake ..
make

The executeable MultiAgentExploration should now be available in build/. CMake also copies the directiories config/ and ẁorld/ into the build directory. Those are required for executing the binary.

For executing the python scripts in the subdirectory scripts/ Python 2.7 and Matplotlib are required. The following dependencies have to be installed via APT:

sudo apt-get install python2.7 python-tk python-matplotlib

The binary of the project can only be executed with the working directory in the build directory, e.g. build/, because of relative paths to config files. Specify with -f <configfile> a config file, which describes the scenario that should be simulated. Those config files can be found in subdirectories of the config/ directory (copied by cmake). -h prints an help text for additional commands.

Example of execution:

cd <build-dir>
./MultiAgentExploration -f config/office/nodecounting-4.yaml

The config files have the follwing pattern config/$TERRAIN/$ALGORITHM-$ROBOTCOUNT.yaml.

To compile an executable with libstage copy the file stage.pc from the Stage build directory to /usr/lib/pkgconfig. Now add the following lines to your CMakeLists.txt:

include(FindPkgConfig)
pkg_search_module( STAGE REQUIRED stage )

include_directories( ${STAGE_INCLUDE_DIRS} )
link_directories( ${STAGE_LIBRARY_DIRS} )

# create an executable and link libstage to each
add_executable( ${PROJECT_NAME} ${PROJECT_SRC} )
target_link_libraries( ${PROJECT_NAME} ${STAGE_LIBRARIES} )

Stage headers can be found in /usr/local/include/Stage-4.1.