Middleware Control Systems Design and Analysis Using Message Interpreted Petri Nets (MIPN)

Abstract

Many distributed frameworks use a message-oriented middleware to interchange information among several independent distributed modules. Those modules make up complex systems implementing basic actions and reporting events about their state. This paper introduces the Message Interpreted Petri Net (MIPN) model to design, analyze, and execute the central control of these middleware systems. The MIPN is a new Petri net extension that adds message-based high-level information communications and hierarchic capabilities. It also contributes to the definition and study of new properties such as terminability for the hierarchy-wide analysis of a system. Special attention is given to the analyzability of the model. Useful relations between the individual properties of each MIPN and the global properties of a hierarchic MIPNs system are extracted through a mathematical analysis of the model. The goal is to analyze each net separately and then build up the properties of the whole system. This results in a great aid for the programmer and optimizes the development process. This paper also shows the actual integration of this new MIPN model in different robot control frameworks to design, analyze, execute, monitor, log, and debug tasks in such heterogeneous systems. Finally, some applications created with this framework in the fields of robotics, autonomous vehicles, and logistics are also presented.

Keywords

• Middleware control systems
• petri nets
• robotics
• distributed control systems
• inter-process communication systems