Modeling of Collision Resolution Algorithm in LonWorks Networks

Abstract

The industrial organization of productive processes have presented a tendency for dispersion and distribution of manufacturing plants based on an increase of resources and functionality of information and mobility technology. In this sense, Local Operating Networks (LON, LonWorks) is the name of one of the leading technologies in sensor/control networking, addressed to a wide range of applications with topology free technology. In spite of its popularity, only a few design tools are available to simulate LonWorks architectures and predict their performance. In this paper, a model of the collision resolution algorithm of LonWorks is described. The approach developed in this work is based on the characterization of LonWorks networks as Discrete Event Dynamic Systems (DEDS), since dynamic behavior is defined through the discrete events and discrete states. The proposed procedure employs techniques that are derived from interpreted Petri net, which has been used as an efficient tool for modeling, analysis and control of DEDS. In this context, the media access control sublayer (MAC) is modeled in different levels of abstraction: a conceptual model which is obtained using the PFS (Production Flow Schema) technique, and a functional model by using MFG (Mark Flow Graph). The MFG abstraction level describes details in a functional form preserving the description activities of upper levels. This procedure allows the structured development of models, facilitating the modeling process of the algorithm specification. The result presented in this paper for a single network segment can be integrated into a global networks modeling, since the proposed procedure decomposes the model systematically according to a hierarchal approach into different modules. In this way, it is straightforward to integrate single segments networks models into complex networks.