Search

In order to improve the ability of real-time ship traffic accident prediction, the support vector machine (SVM) technology was introduced to construct the SVM model. The historical data are used to verify it and results show that the model is effective, and the accident classification accuracy is 82.13%; Compared with BP neural network and Bayesian method, the SVM model has better accident classification accuracy and overall classification accuracy, but the non-accident classification accuracy is lower than both.

According to the complicated calculation process and low practicability of Support Vector Machine(SVM) model in the prediction of ship traffic accidents due to large number of variables, first Random Forest(RF) model is used to rank the importance of variables in the article, then filtered variables is used to construct a new SVM model, finally validity of the new model is verified based on historical data and compared it with the prediction results of the original model. By comparing with prediction results of the original model, results show that the new model is not only relatively simple in calculation but also has better prediction effect, and classification accuracy rate provides 4.12%.

Sensing task allocation is one of the most challenging issues in the field of mobile sensor networks (MSNs). Many existing studies have focused on various aspects of sensing task allocation, such as the relationship between sensing task allocation and the number of participants, the sensing task completion time, and the reward required to complete the task. However, few studies have focused on the relationship between sensing tasks and agent attributes. To address this issue, we first analyze the relationship between the structural characteristics of the mobile perceptron network and the agents. Further, we model the relationship between agents using multi-attribute fusion. Then, according to this model and its preference for sensing tasks, we proposed the sensing task allocation algorithm based on Dempster-Shafer (D-S) theory and multi-attribute fusion(STADMF). Finally, STADMF is compared with three algorithms on large-scale real data sets and a synthetic mobile crowdsensing (MCS) trace. The results show that the proposed algorithm achieves good performance in terms of the similarity and accuracy of task sensing.

Aiming at the problem of network performance degradation caused by the characteristics of mobile ad hoc network, this paper proposes a cooperative packet forwarding algorithm based on game combination and third-party altruism. First of all, we analyze the static game among nodes, and calculate the forwarding probability of nodes in the case of static game through the theorem of Nash equilibrium. Then, for the sake of solving the shortcomings of static game, we introduced the self-game model. Next, this paper puts forward the definition of node’s benefit, which is vague in many studies, and proposes the concept of incentive factor based on third-party altruism to improve node collaboration. Finally, using the static game and self-game, combined with the incentive factor to calculate the node forwarding probability, the data packet cooperative forwarding algorithm is designed. The simulation results demonstrate that the proposed algorithm improves the performance indicators of MANET such as packet delivery ratio, average transmission delay, throughput and routing overhead.

In recent years, sensor multimodal time series data fusion has attracted widespread attention. One of the key challenges is to extract features from multimodal data to obtain shared representations, which combines the characteristics of time series data to further improve prediction performance. To solve this problem, based on Stacked Sparse Auto-Encoder (SSAE) and Long Short-Term Memory (LSTM), we propose a multimodal time series data fusion model called SSAE-LSTM. SSAE mines the inherent correlation features of multimodal data to extract a good shared representation, which is used as the input of the LSTM neural network to perform data fusion processing. Experiments on real time series datasets demonstrate that SSAE-LSTM can obtain a good shared representation of multimodal data to predict the future development trend. Compared with other neural networks, SSAE-LSTM has better performance in Precision, Accuracy, Recall, F-score and so on.

With the popularization of intelligent hardware, wireless sensor networks have led to mobile crowdsensing (MCS) systems, which provide solutions for large-scale and complex urban data collection. Task distribution is the most important part of intelligent hardware applications. MCS can improve the task distribution efficiency by accurately predicting the location of a perceived user for task distribution. This paper proposes a task location estimator based on a variable-order Markov time window sensing (TEMTWS) algorithm. This method is based on time window modeling, and the association between user tasks is established by sensing the historical track data of user execution tasks. First, the task execution frequency and task vector are calculated, and the organizer at each position is selected. To obtain more perceptual users, similarity estimation is performed on the users and organizers within the time window, and users with high relevance are grouped into the same cluster. An experiment is conducted with the Gowalla dataset to verify the algorithm. The results show that the proposed algorithm outperforms the standard Markov K-means algorithm and K-means-GA algorithm in terms of the prediction accuracy.

In order to further improve accident prediction accuracy of real-time ship traffic in waterway, based on ship detector data and traffic accident data collected on two downstream waterways of Yangtze River, important variables were sifted with random forest (RF) model from the initial data of waterway status within 20-40min before the traffic accident, then new Bayesian network(BN)model was established with 4 most important variables combined with Gaussian mixture model (GMM) and maximum expectation(EM) algorithm. Compared with BN model previous studied built with direct initial data, the new models complexity is not only reduced and its prediction effect is improved, with the accident prediction correct rate of 81.29%. The transfer-abilities of two models were also compared, the results show that though accident prediction correct rate of new model is decreased a little, its value is still higher than old model, moreover, new model is much better than old model in terms of non-accident and overall prediction correct rate and their rates are 94.63% and 91.86% respectively.

In order to access sailing risk of real-time ship traffic in waterway and give accident warning in advance, Bayesian network was built by using Gaussian mixture model and maximum expectation algorithm. Based on ship detector data and traffic accident data collected on caoxiexia waterway of Yangtze River, the structure and parameters of Bayesian network were studied and trained, and the segregate was also established. Eight group ship traffic data were adopted to develop traffic accident risk prediction models of BN. Results show that correct rate of accident prediction is 78.13%. At last, a comparative research was done among BN, BP neural network and K nearest neighbor and the results show that the BN prediction model is better than the other two and is a good evaluation method on real-time ship traffic sailing risk.

In order to improve the comprehensive evaluation method on port efficiency, a new general index system is established on the basis of researched results. An integrated evaluation model of efficiency is built on the base of theory of BP network. Based on the efficiency status of port, 10 samples are selected to train BP network and the network model of 7×6×5 in structure is set up. The efficiency of 16 port are assessed with the evaluation model of BP network, the results are basically identical compared with the results of the evaluation model of MDEA.

In cognitive radio networks (CRNs), secondary users need to first discover neighbours and form communication links, referred to as the rendezvous process. Rendezvous between any two secondary users can only be achieved on the same channel. However, the nature of the CRN makes this a challenging problem. Specifically in CRN, not only the network is multi- channel, but the channels available at different nodes may be different. While most of the existing works study pair-wise rendezvous and design channel hopping sequence, in this paper, we focus on the performance improvement of the rendezvous process based on the existing channel hopping sequences with multiple users in CRN. We propose a new strategy, called Group Based Strategy (GBS) to achieve the acceleration, which is flexible to incorporate the existing sequence generation algorithms. Our basic idea is to group the encountered users and schedule rendezvous for them. With the purpose to increase rendezvous diversity, other users or groups can join the group if they get the group rendezvous information. Experiments are conducted to evaluate the proposed scheme. Overall, the performance can be improved by more than 50% under symmetric model or asymmetric model using our accelerating strategy.

This paper addresses the problem of distributed recursive estimation of the state of a dynamical system, using measurements obtained by a network of locally-communicating sensors. Prior work on this problem has mostly focused on the development of methods suitable for very large networks, where multiple iterations of communication between neighboring sensors are used to obtain good asymptotic estimation accuracy. By contrast, we here focus on the case of medium-sized networks (consisting of up to a few tens of nodes), where limited communication is the primary constraint. We propose a distributed estimation method that uses knowledge of the network topology to explicitly model the accuracy of different nodes' estimates, as well as the correlations between them. As a result, the proposed method is able to optimally utilize the measurements and state estimates communicated between neighboring nodes. This method does not require iterations, imposes no assumptions on the structure of the communication graph, and can provide an accurate representation of the accuracy of the produced estimates. We provide Monte-Carlo simulation results that demonstrate that the proposed method outperforms existing algorithms, when all methods use the same amount of inter-node communication.

Recently, mobility-enabled wireless sensor networks have been exploited to improve energy efficiency at the cost of increased latency. In this paper, we propose a Grid-based solution for Real-time Data Gathering application with mobile sinks (GRDG). Like many hierarchical structures, the flexible virtual grid partition scheme based on power control is designed to enhance network scalability, reduce routing hop counts, and ensure balanced energy consumption. Moreover, in order to decrease the transmission latency and cater for real-time application requirement, a local routing tree updating approach is presented. Although the quickly constructed routing tree is not optimal, communication overheads are decreased significantly.

With the increasing number of smartphone users, mobile phone sensing applications have been regarding as a promising paradigm which makes use of the smartphones to access the ubiquitous environment data. In this work, we study the sensing task auction problem where there are multiple tasks and smartphone users. The most significant challenge of this problem is how to design a truthful auction mechanisms, which is crucial for auction mechanism design. Thus, we address this challenge by proposing DATA, which is a truthful double auction mechanism for sensing tasks allocation. Different from the existing designs, we are the first to design double auction mechanism for solving mobile phone sensing problem. Besides, we further take the relationship between the utility of task demanders and the number of users that are assigned to do the tasks into consideration, and assign a set of smartphone users to a winning demander which can maximize the winning demander's utility. At last, we conduct extensive simulations to study the performances of the proposed auction mechanism, and the simulation results corroborate our theoretical analysis.

Effected by illumination and complex background, Haar-like feature values have a large change, and cannot sufficiently represent the face image texture information. By analyzing the distribution of Haar-like feature values, we propose a new type of classifiers called Haar-like intensity feature. Experimental results on some hand-labeled examples and MIT-CMU test dataset illustrate that the AdaBoost algorithm using the extensive features can reduce detection time and make higher face detection rate with fewer simple classifiers.

In this paper, we give an O(NlogN) algorithm to establish a spanning tree rooted at any node of height at most n + ⌈log 2 |F|⌉ + 3 in X n − F, where F is a restricted node set with |F| ≤ 2n − 3 and N = 2n denotes the node number of X n . We also give and analyze the simulation results to apply the reliable unicast and broadcast algorithms in the literature and this paper to some existing bijective connection networks such as CQ n , TQ n , 0-MQ n , and 1-MQ n and general bijective connection networks X n . The simulation results make us conjecture that there would be some special bijective connection networks whose diameters are smaller than the smallest diameter ⌈n+1 over 2⌉ of CQ n , TQ n , 0-MQ n , and 1-MQ n .

The dimensions of twisted cubes in the original definition of twisted cubes are only limited to odd integers. In this paper, we first extends the dimensions of twisted cubes to all the positive integers. Then, we introduce the concept of the set of restricted faulty nodes into twisted cubes. We further prove that under the condition that each node of the $n$-dimensional twisted cube $TQ_n$ has at least one fault-free neighbor its restricted connectivity is $2n-2$, which is almost as twice as that of $TQ_n$ under the condition of arbitrary faulty nodes, the same as that of the $n$-dimensional hypercube. Moreover, we give an $O(N\mbox{log}N)$ fault-free unicast algorithm, where $N$ denotes the node number of $TQ_{n-1}$. Finally, we give the simulation result of the expected length of the fault-free path gotten by our algorithm.

In this paper, we study the restricted connectivity and restricted fault-diameters of locally twisted cubes under the condition that each node has at least one fault-free neighbor. First, we prove that under the condition that if each node of an $n$-dimensional locally twisted cube $LTQ_n$ has at least one fault-free neighbor its restricted connectivity is $2n-2$, the same as that of the $n$-dimensional hypercube. Then, we give an upper bound on the restricted fault-diameter of $LTQ_n$, that is, the restricted fault-diameter of $LTQ_n$ is no more than the fault-diameter of $LTQ_n$ plus 6.

In the field of search, the application of ontology is an important research topic. Introduction of ontology technology in the retrieval system with massive data can make the searching results more comprehensive. However, now days the ontology is constructed by domain experts, and there are a lot of shortcomings, such as complex process, long time for the project, and difficulty to update. Thereby, in this paper, a method of semiautomaticly building ontology is proposed, after synthetically analyzing a variety of methods and techniques about it. The building process which is based on user interests, mines not only the concepts but also the potential relationships between concepts from the texts by the method of concepts clustering. On the basis of such research, an unique patent information retrieval system based on ontology has been completed.

Operation of the JLab IR Upgrade FEL at CW powers in excess of 10 kW requires sustained production of high electron beam powers by the driver ERL. This in turn demands attention to numerous issues and effects, including: cathode lifetime; control of beamline and RF system vacuum during high current operation; longitudinal space charge; longitudinal and transverse matching of irregular/large volume phase space distributions; halo management; management of remnant dispersive effects; resistive wall, wake-field, and RF heating of beam vacuum chambers; the beam break up instability; the impact of coherent synchrotron radiation (both on beam quality and the performance of laser optics); magnetic component stability and reproducibility; and RF stability and reproducibility. We discuss our experience with these issues and describe the modus vivendi that has evolved during prolonged high current, high power beam and laser operation.