For accurately measuring the energy-saving contribution of air traffic management technology on air transportation, this paper proposed a calculation method of fuel consumption in the air traffic control area based on radar tracks. This paper firstly analyzed nine influencing factors, including aircraft type, flight state, true airspeed, and altitude, that could affect aircraft fuel consumption. Taking air traffic trajectory data as input, a fuel flow time series prediction model based on echo state network was built. The predicted approximate error of the model can reach 0.032%, 1.79%, and -1.11% in level flight, climbing state, and descending state, respectively. Due to aircraft weight and missed calibrated airspeed data in radar tracks, a key influencing factors extraction method for fuel consumption based on sensitivity analysis has been further explored. Input parameters of the ESN fuel flow time series approximate model have been simplified reasonably. The Xiamen ATC area was taken as an example, and the total fuel consumption of 1021 flights on a specific day within the Xiamen control area was calculated to be 1044.84 tons. Research results in this paper will construct a technical foundation for measuring air traffic control system performance through implementation of the ASBU plan.

A label propagation algorithm has attracted widespread attention in community detection due to its linear time complexity. However, the traditional label propagation algorithm has a strong problem of randomness and may bring in backtracking during the process of label propagation; the result of finding the community is unstable and of low quality. This essay proposes a circular spread label propagation algorithm (CS-LPA), which takes full account of the structural characteristics of the community, introduces node influence measures, and discovers the potential community through the proliferation of labels that integrate the cyclic structure of social network. Finally, experimental results of real datasets show that CS-LPA not only enhances the stability of community detection results, but also effectively improves the quality of community detection.

The electromagnetic spectrum is an important national strategic resource, and spectrum sensing data falsification (SSDF) is an attack method that destroys the cognitive network and makes it impossible to be used effectively. Malicious users capture the sensory nodes through cyber attacks, virus intrusions, etc., tampering with the perceived data and making the cognitive network biased or even completely reversed. In order to eliminate the negative effects caused by the identification and elimination of abnormal information in the electromagnetic spectrum in multi-user collaboration and to ensure the desired effect, this paper studies and constructs a robust cognitive user evaluation reference system based on improving the performance of cooperative spectrum sensing. The impact of attack behavior on the reference frame is greatly reduced. At the same time, the attacker’s identification and elimination algorithm are improved, and the influence of abnormal data on the perceived performance under the combined effect is eliminated.

The increasing popularity of Internet of Things (IoT) technology has greatly influenced the production mode and life quality of humans. Simultaneously, the security issues of such technology have become a focus of attention. There are many aspects of IoT security issues. In this paper, we propose a framework to solve the problem of network intrusion detection in IoT. First, an intrusion detection dataset named UNSW-NB15 is selected as the research object. Then, the dataset is preprocessed and the feature selection job is accomplished to obtain a suitable subset. After the above steps are completed, a Bayesian model is built according to the K2 structure learning algorithm. The parameters are obtained through the Maximum Likelihood Estimation algorithm. Finally, the testing dataset is inputted for classification. The simulation results show that the system can detect the anomaly intrusion effectively.

Advanced high-power gas turbine is the main research and development direction of future ship power. Three-shaft intercooled-cycle gas turbine engines as the prime mover for marine integrated electric propulsion system have been extensively used in the Chinese navy, and their dynamic performance has attracted many investigators’ attention. This paper extends current research on further improving three-shaft intercooled-cycle gas turbine engine operational performance. Firstly, on the basis of Matlab/Simulink software platform, a nonlinear three-shaft intercooled-cycle gas turbine engine thermodynamic model is set up to simulate the engine dynamic performance. Its power/free turbine shaft speed should be kept constant so as to obtain good dynamic performance control quality for the engine generation set. Under transient loading or unloading operating mode, the power/free turbine shaft speed easily deviates from the nominal value due to operational requirements of increasing or reducing power. In view of the phenomenon that the power turbine is prone to overspeed during the sudden load shedding process, the control strategy of the intercooled-cycle gas turbine power generation module is studied. A dual-mode switching control using fuzzy adaptive PID controller and PI controller and a joint control scheme of high-pressure compressor blow-off controlling are proposed. This provides a control strategy for the actual use of the marine intercooled-cycle three-shaft gas turbine integrated electric propulsion system in the future.

With the development of battery technology, lithium batteries are widely applied to electrical vehicles. The generation of the lithium battery fault has certain complexity and uncertainty, and the quantity of lithium batteries's real-time data test point is low. In addition, the test data is incomplete. Therefore, a fault diagnosis method for lithium batteries is presented based on a fuzzy Bayesian network, and a fault diagnosis model is established combined with fuzzy mathematics and the Bayesian network. The data is fuzzified by fuzzy mathematics to obtain the membership of fault symptoms. The demand of date and computation complexity is reduced by the Leaky Noisy-OR Bayesian network model. If the amount of fault nodes is large, the demand of conditional probability is reduced greatly, from 2n to 2n, by applying the Bayesian network constructed by the model presented above. This method requires less diagnosis time and sample demand, and it has high quality of diagnosis as well as many other advantages. The fault diagnosis of lithium batteries is supported by this method.

XML tree pattern query, also known as Twig query, is the core operation in XML query processing. In the research of the Twig query algorithm, TreeMatch is considered to be one of the best algorithms because it reduces the generation of intermediate results. However, in the core operation getNext of the TreeMatch algorithm, there are many calculations that depend only on Twig mode. This redundant duplicate calculation affects the performance of the TreeMatch algorithm when there are many getNext calls. In order to further improve the algorithm, this paper proposes a Twig query optimization method based on partial evaluation and hot-trace compilation. This method takes Twig mode as an invariant to perform partial evaluation and translates query requests into a Twig query machine instruction sequence. The duplication calculation of the Twig pattern during the query process is avoided, and the process of interpretation of the instruction sequence of the query machine is optimized by using the hot trace compilation technique. The comparison experiment shows that the optimization method based on partial evaluation and hot-trace compilation can increase the efficiency of twig query by 20% to 60%.

Modern processor micro-architecture offers advanced prefetch mechanisms that are designed to effectively hide memory latency and improve application performance. However, pointer-chasing applications employing linked data structures expose a memory latency problem that is difficult to deal with by using hardware prefetchers. It is promising that helper threaded prefetching based on Chip Multiprocessor is an effective method for reducing the memory latency of accesses to linked data structures. In this paper, we first illustrated two L2 prefetchers on Chip Multiprocessor and two different helper threaded prefetching techniques for pointer-chasing applications. Then, we revealed the limitations of L2 prefetchers for pointer-intensive applications after applying two different threaded prefetching techniques. Finally, we optimized the deployment of L2 prefetchers with two different threaded prefetching techniques for pointer-chasing applications. The experimental results indicate that L2 prefetchers’ effectiveness on helper threads depends on the memory access pattern of the targeted applications, and the optimized deployment of L2 prefetchers further improves the performance of pointer-intensive applications.

Data race is widespread in multi-thread programs and can lead to serious failures. To improve the reliability of programs, many race detectors have been proposed. However, most of the detectors use binary instrumentation to detect potential races, imposing higher runtime overhead. Most of the potential races are false positive, which consumes manual effort to identify the harmful races. In order to reduce runtime overhead of identifying harmful races, we propose two concurrent strategies that reduce runtime overhead from the detection potential races stage and the verification harmful races stage. Unlike previous work, the detection and verification races are in one execution. In this paper, the Hadoop distributed system is used to detect the potential races concurrently from the trace, and then the weighted Round-Robin algorithm is used to divide all potential races to multiple cloud servers. Harmful races are verified concurrently in multiple cloud servers by controlling thread scheduling. The experimental results show that our method for identifying harmful races is more efficient. Compared with RaceFuzzer and ReceChecker, the runtime overhead is reduced by an average of 72% and 46% respectively. In addition, a good speedup is achieved in this paper.
Submitted on July 9, 2018; Revised on August 8, 2018; Accepted on September 15, 2018
References: 22

We study the indoor localization problem based on Pedestrian Dead Reckoning (PDR) by analyzing the causes of localization error during pedestrian walking. To optimize the PDR-based localization method, we firstly propose a step-sense indoor localization framework, namely, Stepsense, which can analyze the total acceleration of the accelerometer sensor and obtain the number of pedestrian steps using the peak detection. In the Stepsense framework, the step length is calculated by the difference between the acceleration peak and the trough. The 9DOF (Degree of Freedom) and 6DOF methods are invoked by double-check strategy to make the result of direction estimation more accurate. Secondly, the adaptive error model is used to correct the state of particles in the particle filter, in which map matching and RSS matching are integrated. Finally, both the Stepsense framework and the proposed fusion localization method are examined in detail through experiments.

Real-time navigation, traffic monitoring, broadcasting messages of accidents and entertainment applications require a large number of data sensed by vehicles in the vehicular Ad-hoc networks (VANET). Collection and aggregation of this data is an essential part of implementing these applications. In this paper, we study the maximize aggregation data (MAD) problem in VANET. Based on the idea of greedy algorithm, two different algorithms are proposed to solve the limited communication MAD problem and unlimited communication MAD problem. The core idea of both algorithms is to construct a dynamic routing tree and scheduling the transmission time of each vehicle simultaneously. Real-time traffic information and changeable remaining time are combined to establish the tree. In order to reduce the invalid transmissions, only vehicles that can carry data within the delay constraint or directly forward data to the target node are considered to become relay nodes, aiding in the transmission process of other vehicles. Simulation is performed on the analog trajectories sets, and results show that our proposed algorithms have a higher collection rate compared with other schemes.

K-nearest neighbor (KNN) query is one of the most important query types in spatial databases and have been widely used in intelligent transportation, roadside assistance, and other fields. In order to improve the query efficiency, in this paper we adopted the MapReduce parallel computing framework of the Hadoop large data processing platform and completed the query of K neighbor moving objects by designing Map, Reduce, Combiner, and other functions. Before the start of the query, the road network was divided into pieces, and each fragment was calculated. The final K-nearest neighbor moving objects were obtained by aggregating the calculated results of each slice to realize the parallel optimization of the KNN algorithm based on road network. The experimental results showed that the performance of the parallel KNN algorithm based on MapReduce was better than that of the serial KNN query algorithm in a large-scale road network environment and a larger K value of query requests.

In this paper, we propose an improved TOA model in the indoor three-dimensional positioning of wireless communication base station. By adding an error correction function to the existing TOA location model, the search domain of solutions is narrowed down. In addition, we design an adaptive genetic algorithm to solve the objective function. The simulation results further demonstrate that the proposed algorithm can not only improve the positioning accuracy compared with the existing DTOA, but also has the characteristics of fast convergence speed and strong robustness. The average localization error of the model is only 1.4041m.

The emergence of human-machine collaboration has adapted to the requirements of big data for high performance computing and complex artificial reasoning, which uses the huge Internet user group and cluster to deal with the increasingly complicated data altogether. In this paper, a bipartite graph matching strategy is proposed to solve the problem of how the crowd and the cluster can collaborate effectively to complete the large data task. The Hopcroft-Karp algorithm of bipartite graph matching not only enhances and extends the Hungarian algorithm, but also considers the field of adaptive segmentation tasks, the degree of association, and the evaluation of the background and ability of the crowd to maximize the matching between the crowd and the segmented task group. The algorithm calculates each influence factor after each match and optimizes the next match, making the best match between the crowd and the task. Through the experiment, the accuracy of the task completion is verified to be the highest.

Security and reliability are of vital importance for a remote system to facilitate effective monitoring and management of a wide range of equipment. A distributed secure monitoring system based on blockchain and IoT technologies is proposed in the paper. Firstly, a lightweight security scheme (LSS) is proposed to provide a high level of security with reduced computational and communication costs. Secondly, the architecture of remote monitoring system is presented. An identity authentication method based on the LSS is proposed to authenticate any new node when joining the network. A secure access control method based on blockchain is also proposed to increase the security of the system in a lightweight and scalable manner. Finally, the proposed monitoring system is analyzed with regard to the security against several well-known attacks.

In today’s society, the problems of false academic qualifications and tampering of students’ scores are increasingly prominent, and the social recognition of data such as students’ scores is declining day by day. At the same time, the single performance of a student cannot directly reflect the level of students’ ability. In this paper, the bottom technology of blockchain is studied, and the blockchain technology is applied to the outcome-based education (OBE) of engineering education. Based on the graduation requirements, students’ scores can be converted into specific competency values and recorded in the blockchain. Through the decentralization of blockchain, tamper-proof, and other characteristics, a social consensus is formed. At the same time, this paper designs an association code, which is more convenient to learn how to find the records in the output blockchain than the traditional way of tracking the account balance when querying the bitcoin records.

Electronic voting has been widely used in modern democratic elections in recent years. Electronic voting has been a popular issue with cryptography technique because of the importance of voting results. There are many requirements for a secure electronic voting scheme. It is difficult for existing solutions to solve these incompatible requirements. Blockchain technology provides a new solution to electronic voting schemes. In this paper, we proposed an anonymous electronic voting scheme based on blockchain. This scheme effectively protects the privacy of voters. Unlike previously proposed blockchain electronic voting schemes, our scheme can be used in boardroom voting circumstances and allow voters to abstain from voting. More significantly, our scheme uses a blind signature and ring signature to eliminate “double-voting” behaviours. The results of the voting can be self-tallying by any nodes of our network because all ballots are stored in the blockchain bulletin board. At the end of this paper, we analyze our scheme with the requirements of a secure voting scheme.

The multi-objective software module clustering problem (MOSMCP) divides the complex software system into subsystems to obtain a perfect structure, which is based on the relations between modules to meet the conflicting software refactor objectives as much as possible. The modularization quality (MQ) and reverse edges number between clusters are considered as evaluation objectives, and a novel particle swarm optimization (PSO) with reverse edge, called REPSO, is proposed. First, the module dependency graph (MDG) in software system under clustering is constructed, and then the multi-objective particle swarm optimization (MOPSO) is improved to cluster the MDG. The exploring approach is used to update the particle locations. Four typical open source projects for module clustering are selected to verify the effectiveness of the REPSO. The laboratorial results prove that the REPSO algorithm is very effective and stable, and the diversity of the optimal solution is good. The REPSO algorithm provides an efficient engineering method for MOSMCP, which enhances the software structure and maintainability.

DTN enhances message transmission success rate and reduces message forwarding delay by increasing the number of message copies. However, a large number of redundant copies will cause nodes to be congested, which decreases the resource utilization rate of networks, resulting in low overall network performance. To address this issue, we propose a congestion control method NSS-CC, which is based on node storage status. According to the storage space utilization rate of nodes, the storage status is categorized into three states: Normal, Semi-Congested, and Congested. The node adjusts its own congestion control mechanism according to its own congestion state, in order to balance the load of the node effectively, and judges whether to receive the message by defining the Degree of Willingness (DoW). The simulation results show that NSS-CC is superior to congestion control algorithms such as DO, DF, DY, and DL.

To address the Multi-Objective Test Case Prioritization (MOTCP) problem, an Epistatic Particle Swarm Optimization (EPSO) algorithm is presented. The epistasis in biology is introduced into the new algorithm, and the particles are updated based on the crossover of Epistatic Test Case Segment (ETS) in the test case sequence. The average coverage percentage of program entity and effective execution time of the test case sequence are set as two objective fitness functions in EPSO. The experiment selects four typical open12 source projects as benchmark programs. We adopted Average Percentage of Branch Coverage (APBC) and Effective Execution Time (EET) as objective fitness. The four classical Java testing projects results show that the EPSO is more effective and more diverse than single-point PSO and order PSO. The EPSO algorithm efficiently solves the MOTCP problem by promoting early detection of software defects and reducing software testing costs in regression testing.

This paper proposes a novel approach called ITE to extract web data records in a fully automatic way. The approach effectively utilizes the tag index information in different layers of the HTML DOM tree and abstracts the concept of index tree together with its repetitiveness and consecutiveness, which can characterize the key structural information in a web page. The concept of repetitiveness indicates the structural similarities among data records, and the concept of consecutiveness represents the sequential features of multiple records. Then, the complex DOM tree can be compressed to a set of index trees based on these concepts. We also provide a series of properties as theoretical support. The extraction process is divided into three steps, namely, repetitiveness discovery, consecutiveness discovery, and index tree merging. To handle data field missing, multiple record roots, and other complicated situations, we propose a digital sequence similarity measurement and a hierarchical clustering approach to find the repeating patterns. Then, data records are identified based on the consecutiveness discovery method, and the data blocks containing full data records are restored by merging the index trees. Experiments demonstrate the effectiveness and efficiency of the proposed approach. It outperforms existing classic work in accuracy and has a satisfying execution time, which means it is applicable to large datasets. The time complexity is linear to the number of leaf nodes in the DOM tree of a web page.

Maritime traffic pattern is very important for intelligent ship collision avoidance applications, as it can help provide decision support to avoid ship collision accidents and reduce casualties. There has been a large amount of Automatic Identification System (AIS) data saved by ports or management departments. If these data can be processed and analyzed scientifically to provide an early warning with appropriate collision avoidance measures, injuries or more serious results from maritime traffic may be reduced or eliminated. Our focus is to synthesize ship behaviors of interest in a clear and effective way based on automatic preprocessing and analyzing original static AIS data. One improved DBSCAN algorithm is first called to reduce the data scale and discover important data points. Then, from the perspective of Own ship, seven patterns including course change and speed change are defined to be discovered. For each special pattern, the space collision risk DCPA (distance to closest point of approach) and time collision risk TCPA (time to closest point of approach) at the beginning time and ending time are computed to confirm its situation as heading on, crossing, or overtaking with other ships in sight of one another. This unsupervised learning approach will help discover traffic pattern knowledge in current trajectories and provide decision support for future route design or anomaly analysis.

In practice, some attributes meet a unique constraint: each entity has a unique value for the attribute. A deep web entity identification method was presented to solve problems of data error correction, uniqueness constraint enforcement, and local data fusion in deep web data integration. The method transformed the entity identification phrase to a k-partite graph clustering problem, considering both similarity and association of attribute values. Moreover, it performed global record linkage and data fusion simultaneously and could identify incorrect values and differentiate them from correct ones at the beginning. Experimental results demonstrate the high precision and scalability of our method.

This paper proposes a new highlights event detection method for basketball videos. The support feature of each highlight is firstly found using the concept lattice clustering technology according to the audio-video features and middle level semantic features defined in this thesis. Then, the support features are weighted to construct the affective arousal feature. The audio shots are processed to obtain the whistle shots features using the whistle shots detection method defined in this thesis. The affective arousal feature and the whistle shots features are combined as the input. An effective HCRF (Hidden Conditional Random Field) is constructed to realize highlight detection of basketball shooting and fouls. Experimental results show the effectiveness of the proposed method.

Information overload is a key issue of the current network information retrieval, and a personalized recommendation with special information filtering methods is an important way and means to solve this problem. Based on the analysis of the common methods used of personalized recommendation, the architectural design of the personalized recommendation is proposed on the cloud computing platform. Then, combined with the specific issues of employment recommendation, this article proposes an optimized algorithm of Mahout distributed personalized recommendation based on content and items. Compared with the current single target recommendation algorithm, this algorithm is more efficient with a good practical significance and reference value.

Content-Centric Network (CCN), one of the current research hotspots, is known as the network architecture with the best development prospects. As an important part of CCN, the performance of cache decision strategy directly affects the whole performance of CCN. A hierarchical caching decision algorithm (HCDA) is proposed to solve the problem of existing cache decision strategies in CCN. The strategy grades contents and nodes in network topology so that the content can be hierarchically cached. Doing so solves the redundancy caused by LCE (Leave Cache Everywhere). Simulation results show that compared with LCE and ProbC (Probabilistic Caching), HCD effectively raises the cache hit ratio and reduces the cache content redundancy; thus, the user requests hop is reduced.

Superword level parallel (SLP) algorithm is an automatic vectorization method that is suitable for the applications including parallel codes. Existing SLP algorithm could not efficiently deal with the applications that contain few parallel codes. In the present study, a new compile framework based on the improved SLP algorithm is presented. The framework contains three phases: isomorphic processing for isomeric statements, establishment of superword statements, and data layout optimization. Firstly, isomeric statements with similar instruction in the codes were transformed to isomorphic statements by the improved SLP algorithm. Secondly, the superwords reuse patterns were obtained before making the optimization decisions from a global point of view. Finally, data layout optimization was combined for further performance improvement. The experimental results indicated that the optimization of the compile framework was better than existing SLP algorithm.

Polling control system is widely used in production and life including time-sharing computer systems, industrial control, communications, and computer networks. The article proposes a new polling control system based on Busy/Idle queues, which sorts normal polling queues into Busy Queues (BQ) and Idle Queues (IQ) according to if there are customers in the queue. Then, BQ is served by a Gated access policy and IQ keeps a sleeping state until it is woken up by arriving customers. Moreover, parallel scheduling is used to save switch-over time. We build a system model using the embedded Markov chain, probability mother function, the throughput, cycle time, mean queue length and mean waiting time of significant system characteristics. Theoretical calculated values are approximately equal to the simulated values, indicating that the new system is correct and achieves a better performance than the traditional polling scheme.

To enhance the personalization of an e-learning system, an automatic approach of exercise recommendation that is driven by learning objective is proposed. Firstly, the formal models about knowledge points, exercises and their relations are presented based on a course knowledge tree. Then, a computing method is proposed to constantly and automatically update learning objectives in the learning process. According to the learner’s learning state, an approach is proposed to accurately describe the learner’s learning needs. In order to realize the personalization within the e-learning system, three kinds of influencing factors, including learning objective, the grasp state of knowledge point and learner’s answer preferences, are taken into account for the exercises recommendation. A running example is analyzed to demonstrate the feasibility and validity of the proposed approach for recommending exercise to a complete learning objective in a rapid manner.

Hyperspectral image has a large amount of data and complex structure. The distributed storage of massive remote sensing data is a hot topic today; however, the current research mostly separates the image pixels and metadata, resulting in poor system cohesion and poor data access performance. At the same time, the needs of various upper-level remote sensing algorithms are not fully considered, which makes the system less available. In view of the above problems, this paper presents a distributed image storage model based on HDFS, which stores the entire image data model in a structure to improve the system cohesion, and provides a flexible data blocking strategy for upper-level applications to meet a variety of data access needs. The comparison experiments show that the storage model has better access performance than the existing schemes.