Reactive, model-based monitoring in RFID-enabled manufacturing

Radio frequency identification (RFID) technology is widely applied throughout the manufacturing lifecycle to track important objects. It is imperative to establish a reactive monitoring system to convert large volumes of RFID data into meaningful information. Two major challenges must therefore be addressed: modeling complexity and monitoring efficiency. When the relationship between high-level “untagged” elements (business processes, indicators) and low-level “tagged” elements (persons, objects) becomes intricate, it will become difficult to define, maintain and reason by only using traditional rule-based event processing approaches. In order to resolve this problem, the paper proposes a reactive, model-based approach to monitor important elements by estimating their most likely states according to RFID information and a constraint-based model. The approach contains two phases. (1) The model-based programming phase formalizes RFID information and flexibly embeds them into the model that specifies the monitoring rules. (2) The model-based monitoring phase interprets the model and estimates desired states efficiently. Therefore, the RFID-enabled manufacturing units are capable of responding in a timely manner to unexpected disturbances or other important issues.

► We provide a model-based framework for reactive monitoring in RFID-enabled Manufacturing.
► We integrate rule-based event processing and model-based probabilistic reasoning.
► We separate the event-based detection and the constraint-based reasoning and use appropriate approaches respectively.
► We normalize the RFID event processing lifecycle.
► We define four types of events at different levels: TREs, simple events, complex events and RFID observations.