The Research On Complex Event Processing For Large-scale CPS

Cyber Physical System(CPS)has overlaid every application areas.How to manipulate information of CPS application is the key issue for CPS widely used,and CPS middleware can deal with this information effectively.Complex event processing(CEP)is one of the core tasks of middleware for CPS.Event stream of CPS applications has many features,such as heterogeneous,distributed,mass and uncertainty,et al.In the practical application of the CPS,due to noise,sensor error or wireless communication techniques and other reasons,uncertain complex event processing has become very important.In order to deal with large-scale probabilistic stream,this paper proposes an effective uncertain complex event stream processing.It is based on an existing stream processing engine RFID Complex Event Detection Algorithm(RECDA),it uses the event probabilistic model for probability calculation.In the design of matching tree,we consider optimizations to reduce the multi-query cost with Related Query Table.The experiment shows that the method is effective in handling multi-query uncertain event stream applications.The data produced by internet of things has the characteristic of big-data and those data can hardly be processed by present data processing technology.As the key part of Internet of things,Complex Event Process(CEP)has two characteristics of big data:quantity,complexity and has the need of on-time processing.Context-awareness is an important feature of CEP engine.In this paper a high-performance distributed context-aware CEP architecture and method is proposed for internet of things.The method uses fuzzy ontology to create query model to support event query of uncertainty and fuzzy.Based on fuzzy ontology query and similarity based distributed reasoning,complex event query plans are generated and context-aware queries are rewritten into context independent sub-queries.The sub-queries are optimized and executed parallel based on data partition.The experimental results show that this method can support fuzzy context in CEP and have better performance and scalability than other methods.