Contextual merging of uncertain information for better informed plan selection in BDI systems

Sensor information (e.g. temperature, voltage, etc.) obtained from heterogeneous sources in SCADA systems may be uncertain and incomplete, while sensors may be unreliable or conflicting. To address these issues we apply Dempster-Shafer (DS) theory to correctly model the information so that it can be merged in a consistent way. Unfortunately, existing merging operators are not suitable for every situation. We adapt a context-dependent strategy from possibility theory where we determine the context for when to merge using Dempster´s rule of combination (i.e. for low conflicting information) and then resort to Dubois and Prade´s disjunctive rule to merge information which is highly conflicting. We demonstrate the suitability of our approach with a scenario of a smart grid SCADA system modelled using the Belief-Desire-Intention (BDI) multi-agent framework. In particular, we use the notion of epistemic states to model combined uncertain sensor information for better informed selection of predefined plans.