Granular Computing Application to Hybrid Systems Control

Rough sets are used to define granule of information from data of a sensors network. The abstraction is used to generate a meaning from uncorrelated events with appropriate definitions of dependency. Such granules of information are used to abstract states of a system under control. To have system control means to have a control model. The non linearity and the complexity of the environment bring naturally to the application of piecewise-affine hybrid control system approach. The approach simplifies the granules definition and takes into account the relative significance of the available information (discrete and continuous). The report presents a new approach to integrate event calculus with the control theory of hybrid systems using the glue of rough sets.