DSAAM : A time management library for simulation middlewares

This a proof of concept for a proposed time management scheme for building decentralized simulations.

Contrary to HLA or DIS, the time management scheme is completely decentralized.

The library is implemented in C++ and Python for the ROS middleware, as well as for intra-process communication between threads.

Requirements

C++

A C++14 compatible compiler

Python

Python>=3.5

For the tests:

• numpy
• matplotlib

ROS

ROS>=kinetic with rospy and roscpp packages

Installing

C++

mkdir build && cd build && cmake .. && make install

Python

Use the provided setup.py script, for help see: python3 setup.py install --help

Tests and examples

A synthetic example has been implemented in both python and C++, for the bare intra-process communication method as well as for the ROS middleware

Intra-process communication between threads

Python

To run the example (with graphical output): python3 test/python/test_nbody.py

It can also be run as a test using noesetests3: PYTHON3_INTERPRETER=python3 nosetests3

C++

Building the example: mkdir build && cd build && cmake .. -DBUILD_TEST=ON && make build_tests

Running it (no graphical output): ./nbody_threads

ROS

To build the ROS workspace: mkdir build && cd build && cmake .. -DBUILD_TEST=ON -DBUILD_ROS && make build_tests

To run the example (from the project root folder): source build/ros/devel/setup.bash && test/ros/src/dsaam_nbody/scripts/test_nbody.py

The ROS example mixes python and C++ nodes.