ROS Framework for Distributed Control of Networks of Dynamical Systems

In this paper we present a software package developed to accommodate flexible and efficient modelling and simulation of networks of dynamical systems with distributed control structures. The package is implemented within the Robot Operating System (ROS), which is becoming a standard software implementation tool in robotics, mechatronics and wider. The developed package is characterized with the following features: implementation of distributed controllers with both off-line predefined control laws in closed form and with on-line iterative solving of an optimisation problem (distributed model predictive control); modular structure which allows the user to easily modify the subsystems in the network, as well as the number of subsystems; allows for straightforward implementation of controllers onto real physical systems.