Complex Event Processing(CEP) is an effective approach for real-time event stream monitoring,processing,analyzing and mining.Core concepts and basic elements of CEP are introduced.Major areas that utilizing CEP technology are reviewed,such as Internet of Things(IoT),cloud computing,grid computing,Business Process Management(BPM) and computation finance.Based on these applications,cloud based distributed parallel CEP technologies are introduced.By analyzing the requirements of future stream processing systems,it is concluded that such cloud based CEP technology is an important direction for the development of future CEP systems.Potential new areas that require such cloud based CEP system for complex stream event processing are explored,such as power system intelligent control,instrument control and intelligent measurement.Guidance can be provided for the utilization of the CEP technology for now and future.
For the secure issues of outsourced data in cloud storage,a new method of data integrity verification is proposed,which can support secure deduplication and public verification.The method combines the advantages of Proofs of Ownership(POW) and Proofs of Retrievability(POR),and achieves client-side secure deduplication and data integrity verification simultaneously,in which data block,random sampling and dynamic coefficient are used.By introducing bilinear pairings and erasure codes,users can infinitely verify whether the data is intact in cloud storage.If the data is damaged,users can repair it with erasure codes.To achieve privacy preserving,the technology of random masking is introduced,which can hide the information of users’ data effectively.Analysis results show that the proposed method not only can assure the security and integrity of data stored in cloud storage,but also can efficiently reduce the computing cost and communication cost.
Aiming at the real-time requirements and easy upgrode feature for high-speed signal acquisition and processing in Acoustic Doppler Current Profiler(ADCP),a hardware and software co-design method based on Field Programmable Gate Array(FPGA) is proposed.This paper expounds the principle of current profiler and chooses the FPGA as a signal processor.Using Verilog HDL language to describe some modules which are easy to use hardware to implement,such as synchronous data acquisition,filtering and complex correlation calculation.This paper chooses the FPGA’s MicroBlaze soft core as the CPU system,operating process control,determining branch and invoking hardware modules to control echo signal of acquisition,processing and storage.Experimental results show that the real-time feature of FPGA signal processing meets the requirement of the system,and it has high accuracy.
Existing service discovery methods only consider atomic service or service composition,which cannot meet the increasing demands of users.Aiming at this problem,this paper proposes a new Web service matching framework based on model transformation.In the process of service discovery,this framework comprehensively considers the execution process information of composite Web services information and Quality of Service(QoS) requirements.It designs a whole Web service composition matching process:it transforms UML model into process model,then transforms process model into process graph and calculates the similarity of two combined Web services.Experimental result shows that the proposed framework can distinguish the different combined Web services and improve the accuracy of service discovery.
Web proxy server cache can reduce network congestion in a certain extent,and it can also reduce server load and user’s access delay.However,the Web proxy cache is just passable in the cache hit rate and byte hit rate,cannot play very well to accelerate network request response effect.Combining supervised learning method,this paper tries to classify the Web log data using Tree Augmented Naive Bayes(TANB) classifier,predicts the Web object,and proposes a new cache strategy with the regularly used Least Recently Used(LRU) algorithm.Experimental results show that TANB classifier is superior to the naive Bayes and BP neural network classifier in the precision and recall index.And compared with LRU algorithm,optimized cache strategy cannot only improve the cache efficiency,but also effectively improve the request hit rate and byte hit rate of Web proxy cache.
The existing methods of constructing Connected Dominating Set (CDS) have some drawbacks,such as redundant steps,much more energy consumption,and not adapting to the changes of dynamic network topology.So this paper proposes an improved algorithm called Energy Efficient Algorithm of Constructing a Connected Dominating Set(EEIA_CDS),which can quickly construct a CDS that is adaptable to the tiny movement of nodes just with a single phase.It simplifies the procedure of construction and reduces the energy consumption.Furthermore,this algorithm takes the additional coverage and remaining energy of sensor nodes into consideration while choosing domination nodes.So it prolongs the survival time of backbone network and avoids energy consumption of frequenting construction backbone network.Simulation result shows that compared with the EEIA_CDS,Flooding algorithms the backbone construction expenditure of the proposed algorithm is reduced by about 31%~46% and the occurrence probability of the BSP is reduced by about 52%~67% while network’s lifetime is increased by 35.5%.
Aiming at poor communication problems of Vehicular Ad Hoc Network(VANET) in the adverse road conditions,this paper proposes a routing optimization algorithm which uses the Global Position System(GPS) positioning system and digital map device to retrieve information such as the speed,movement directions and relative positions among different vehicles.The most theoretically reliable communication path for message transfer is chosen for messaging.When the optimal communication transfer path is hindered by the external terrain environment,the backuq copy fallback mechanism is used to avoid communication obstacle,which can improve communication reliability.Simulation results show that the algorithm,compared with Ad Hoc On-demand Distance Vector Routing(AODV) protocol,has better performance in the success rate of packet delivery,the number of linking disconnection and route discovery frequency,and it can ensure the communication quality of VANET.
Most traditional object detection approaches extract features manually,which can hardly detect specific objects in complex High Resolution Remote Sensing Imagery(HRRSI).For solving the object detection in HRRSI,a Multi-Structure Convolutional Neural Network(MSCNN) model is constructed to learn object features automatically.Four CNNs with different network structures are designed to extract features from low-level,mid-level to high-level with respect to various scales by considering the size and as well as the number of convolution filter in the convolution layers and the number of layer.Then,the outputs of the four CNNs are concentrated and put into a BP network for training a classifier.This paper uses the sliding-window method to search the object.Experimental result on airplane detection in HRRSI shows that MSCNN has obvious advantages than single-structure CNN.It not only reduces the false alarm rate by 3%,but also improves the recall rate by 6%.Experimental result on oil tank detection further shows that MSCNN can be used in the detection of the other objects in remote sensing imageies.
Characteristics of human motion are always diverse and complex.The intensity of some actions is quite different at different stages of the movement,but the existing methods don’t take this factor into account when evaluating the similarity of actions,which makes the evaluation results not accurate enough.To address this problem,this paper extracts key frames of four component time series of the most violent joint from reference action sequence by multi-scale Faber-Schauder wavelet interpolation.The four groups of key frames are merged and a threshold is set to exclude the key frames of high similarity.Dynamic Time Warping(DTW) method is used to match the reference action and contrast action to get key frames of the contrast action.The motion similarity score can be calculated by normalizing the average distance between key frames of the two actions.Experimental results show that the proposed method can achieve a better evaluation than other methods and the evaluation result is also better for some similar actions.