摘要：Aiming to solve the problems of big combination number of possible events and high memory consumption and low detection efficiency during the detection process of Naive method for multi-probability RFID event streams, a new complex event detection method based on NFA-DAG (Nondeterministic Finite Automaton-Directed Acyclic Graph) is presented for multi-probability RFID event stream in this paper. The achievement of our proposed algorithm lies that we use the union of NFA and DAG to detect multiple-probability RFID event stream, as a result, it can effectively improve those problems above existed in Naive method. The simulation results show that our proposed scheme based on NFA-DAG in this paper can greatly reduce the combinations number of possible events, lower memory consumption and improve event detection efficiency in detecting multi-probability RFID event streams compared with Naive method, without degrading detection quality.
