This repository includes the implementation of supervisors for discrete event systems againt several types of attacks.

• Python (>=3.5)
• OpenFst(>=1.6.1) with pywrapfst module

The examples in this repository also require the following libraries for visualization.

• Jupyter (>=1.0)
• Graphviz(>=2.36)

You can install the OpenFst library from the Ubuntu repository:

sudo apt install libfst-dev

and you can install the Python wrapper for OpenFst using pip. The version openfst==1.6.1 is currently the only version compatible with libfst-dev.

pip3 install openfst==1.6.1

You need to compile the OpenFst library first with --enable-far option.

curl -O http://www.openfst.org/twiki/pub/FST/FstDownload/openfst-1.6.9.tar.gz
tar -xzf openfst-1.6.9.tar.gz
cd openfst-1.6.9
./configure --enable-far
make -j4
sudo make install

After that you can simply install the Python wrapper using pip.

pip3 install openfst

You can download and install the standalone Python wrappers for 64-bit Windows from the following links:

After that you can simply install the Python wrapper using pip.

pip3 install pywrapfst-1.6.9-cp36-cp36m-win_amd64.whl

Our examples in this repository require Graphviz and Jupyter for visualization. You can install Jupyter using pip. Example:

pip3 install jupyter

You can find detailed installation instructions for Graphviz here.

To install the current release:

git clone https://github.com/alperkamil/arsc.git
cd arsc
pip3 install .

You can synthesize the attack-resilient supervisor for the plant P, the desired language MK, the sensor attacker As and the actuator attacker Aa using the function arsc.supervisor(MK,P,As,Aa) where MK,P,As and Aa are all pywrapfst.Fst objects. For example,

$ python

>>> import arsc
>>> MK,P,As,Aa = arsc.example()
>>> MK
<vector Fst at 0x2a0a94178f0>
>>> S, controllable = arsc.supervisor(MK,P,As,Aa)
>>> S
<vector Fst at 0x2a0a9417c00>
>>> controllable
True