Search

In complex underwater environment, the outputs of Doppler velocity log (DVL) often contain outliers induced by external disturbance or soft failure, which will degrade the performance of the traditional Kalman filter based INS/DVL integrated navigation system. To address this problem, a novel robust Kalman filter with outlier-contaminated measurement noise modelled as elliptically contoured distributed is proposed. Firstly, an analytical update form of the joint posterior probability density function is derived via introducing an unknown parameter. Then the fixed-point iteration method is exploited to achieve an approximate estimate of the unknown parameter, based on which the required joint posterior probability density function is analytically updated. Finally, simulation results of INS/DVL integrated navigation are used to verify the validity and advantages of the proposed method.

To address the fixed-interval smoothing problem for a linear state-space model with non-stationary heavy-tailed measurement noises, a robust Student’s t-based Rauch-Tung-Striebel smoother is proposed in this paper. The measurement noise is modelled as a non-stationary Student’s t-distribution, which is written as a Gaussian hierarchical form by introducing Gaussian and Gamma random variables. The time-varying state trajectory, Student’s t-distribution parameters and auxiliary random variables are jointly estimated by using the variational Bayesian method. Simulation study shows that the proposed smoother outperforms the existing up-to-date smoothers in the face of non-stationary heavy-tailed measurement noises.

In recent years, lightining faults occur frequently with the rapid development of power grids and the increase of strong convective weather. Aiming at the problem that the calculation results of the existing calculation method of lightning strike rate of the transmission line are quite different from the actual operation situation, this paper proposes a lightning strike risk assessment model for the transmission line based on the multi-support vector machine combined classifier. This paper evaluates the risk of lightning strikes of the tower by using the actual lightning operation data of 500kV overhead transmission line tower and counting the probability output of all SVM sub-classifiers. An adaptive genetic algorithm is proposed to optimize the parameters of the SVM classifier, which effectively improves the problem of low classification accuracy of the model due to the large penalty parameter value in the conventional genetic algorithm. The results show that the evaluation results of multi-SVM combined classifier can greatly reduce the number of towers that are misjudged as high risk under the principle of evaluating the conservativeness of lightning strike risk on transmission lines.

As one kind of energy storage (ES) applications, ES can respond to electricity prices and help electricity users obtain economic benefits. In detail, by storing electricity during low price period and releasing power energy during high price period, ES can obtain price arbitrage or lower the energy cost for power consumers. However, among the existing ES arbitrage methods, most of them should build sophisticated modeling considering the power prices and several other ES constraints, including battery charge and discharge constraints, State-of-charge (SOC) constraints, state changes constraints, etc. The calculation is comparatively complicated and needs to be solved rely on some intelligent algorithms or professional computing tools. In this work, a fast calculation method supporting arbitrage under Time-of-Use (TOU) price for ES is proposed. The electricity price signal and ES operation factors are comprehensively considered, with a heuristic and adaptive principle suitable for ES response under electricity prices proposed. A simulation case towards an ES station considering a regional power grid is performed, presenting the proposed approach can help ES determine the optimal charging/discharging scheduling options. The calculation method needs very few computational efforts and is easy to be implemented, thus offering an effective means for ES to obtain price arbitrage.

Stray current, as a special current, is common in the earth, and it will have a certain impact on the grounding grid of the power system. Based on CDEGS modeling, this paper studies the distribution of stray current and its influencing factors. In addition, the stray current in the soil not only causes corrosion to buried metals, but also affects the cathodic protection system of buried metals to varying degrees. It is proved through simulation that a certain level of stray current will have a serious impact on the sacrificial anode cathodic protection method, but the external current cathodic protection method has little effect. Therefore, it is recommended to use the external current cathodic protection method in the area of stray current interference. Combined with an engineering example in a certain area, a corresponding external current cathodic protection scheme is designed according to the analysis of the factors affecting the stray current, which greatly reduces the corrosion rate.

In large-scale multi-agent systems, the existence of a large number of agents with different target tasks and connected by complex game relationships causes great difficulty for policy learning. Therefore, simplifying the learning process is an important issue. In multi-agent systems, agents with the same target tasks or attributes often interact more with each other and exhibit behaviors more similar. That means there are stronger collaborations between these agents. Most existing multi-agent reinforcement learning (MARL) algorithms expect to learn the collaborative strategies of all agents directly in order to maximize the common rewards. This causes the difficulty of policy learning to increase exponentially as the number and types of agents increase. To address this problem, we propose a goal-based hierarchical group communication (GHGC) algorithm. This algorithm divides the agents into different groups, and maintains the group’s cognitive consistency through knowledge sharing. Subsequently, we introduce a group communication and value decomposition method to ensure cooperation between the various groups. Experiments demonstrate that our model outperforms state-of-the-art MARL methods on the widely adopted StarCraft II benchmarks across different scenarios, and also possesses potential value for large-scale real-world applications.

One of the toughest challenges in the multi-agent deep reinforcement learning (MADRL) is that when the opponents’ policies change rapidly, the collaborative agents can’t learn well to respond to the opponents’ policies effectively. This may lead to a local optimum w.r.t. the learned policy of the collaborative agents may be only locally optimal to the opponents’ current policies. To address this problem, we propose a novel algorithm termed Friend-or-Foe Deep Deterministic Policy Gradient (FD2PG), in which the cooperative agents can be trained more robust and have stronger cooperation ability in continuous action space. These collaborative agents can generalize easily and respond correctly, even if their opponents’ policies alter. Inspired by the classic Friend-or-Foe Q-learning algorithm (FFQ), we introduce the idea of minimizing the foes and maximizing the friends into the centralized training distributed execution framework, multi-agent deep deterministic policy gradient algorithm (MADDPG), to enhance collaborative agents’ robustness and cooperativity. Besides, we introduce a Minimax Multi-Agent Learning (MMAL) method to explore two special equilibriums (the adversarial equilibrium and the coordination equilibrium), which can guarantee the convergence of FD2PG and improve optimization. Extensive fine-grained experiments, including four representative scenario experiments and two scale-performance correlation experiments, were conducted to demonstrate the superior performance of FD2PG comparing with existing baselines.

Deep reinforcement learning has made significant progress in multi-agent tasks in recent years. However, most previous studies focus on solving full cooperative tasks, which do not perform well in mixed tasks. In mixed tasks, the agent needs to comprehensively consider the information provided by friends and enemies to learn its strategy, and its strategy is sensitive to the received information. There is a great necessity to efficiently learn information representation for mixed tasks. To this end, we present an approach that conducts information representation learning for multiple agents using hierarchical attention mechanism. Our approach adopts the framework of centralized training and decentralized execution. It applies hierarchical attention to centrally computed critics, so critics process the received information more accurately and assist actors to choose better actions. The hierarchical attention critic uses two different attention levels, the agent-level and the group-level, to assign different weights to information of friends and enemies respectively and then summarize them at each time-step. It can achieve more effective and scalable learning in mixed tasks. In addition, our approach uses recurrent neural networks that process sequence input information more efficiently. Experimental results show that our approach is not only applicable to cooperative environments but also better in mixed environments. Especially in the predator-prey task, our approach receives twice as much reward as baselines.

The failure rate is a reliability evaluation parameter based on probability in the power system. It is very important to calculate the reliability index and the equipment fault loss through the failure rate for the planning and operation of the power system. A failure model of large transformer failure rate is proposed. Firstly, the failure rate model of the transformer is established, and the current failure rate and running state of the transformer are obtained. Secondly, a multi-state Markov failure rate prediction model is established to predict the real-time failure rate of the transformer. Finally, the proportional failure rate model of aging failure process to correct the current state of the transfer rate of transformer, and then on the basis of a more accurate transformer real-time prediction failure rate. Example analysis results show that the multi-state Markov failure rate prediction model can be modified by the aging failure model, and the real - time failure rate can be predicted more accurately.

In order to alleviate the impact of grid-connected new energy micro-grid including wind power and photovoltaic on the safe operation of the system, a micro-network optimization scheduling strategy considering energy storage system under time-sharing electricity price is proposed. The investment cost, pollution penalty cost and main network purchase cost are considered respectively, and the micro-network main network joint operation optimization model is established. The simulation results are used to analyze the optimization results under different scenarios. The results of the example show that the proposed strategy and model can effectively realize the micro-network optimization scheduling, effectively reduce the annual operating cost of the micro-grid including electric vehicles and storage batteries, and ensure the joint safe operation of the micro-network and the main network.

Code virtualization built upon virtual machine (VM)technologies is emerging as a viable method for implementing code obfuscation to protect programs against unauthorized analysis. State-of-the-art VM-based protection approaches use a fixed set of virtual instructions and bytecode interpreters across programs. This, however, exposes a security vulnerability where an experienced attacker can use knowledge extracted from other programs to quickly uncover the mapping between virtual instructions and native code for applications protected under the same scheme. In this paper, we propose a novel VM-based code obfuscation system to address this problem. The core idea of our approach is to obfuscate the mapping between the opcodes of bytecode instructions and their semantics. We achieve this by partitioning each protected code region into multiple segments where the mapping of opcodes and their semantics is randomized in different ways in different segments. In this way, each bytecode instruction will be translated into different native code in different sections of the obfuscated code. This significantly increases the diversity of the program behavior. As a result, the knowledge of bytecode to native code mappings obtained from other programs will be less useful when targeting a new program. We evaluate our approach on a set of real-world applications and compare it against two state-of-the-art VM-based code obfuscation approaches. Experimental results show that our approach is effective, which provides stronger protection with comparable runtime overhead and code size.

The large spread in recent published $G$ values suggests that there may be some undiscovered systematic sources in these experiments. With the aim of identifying possible errors and obtaining a more reliable $G$ value, our group has performed two independent methods simultaneously. Besides the time-of-swing method which we have worked on for more than 20 years [1], [2], another determination of $G$ with the angular acceleration feedback method was carried out at the same laboratory. In the AAF method, two coaxial turntables are used to rotate the torsion pendulum and the source masses individually. With the help of a feedback control system [3], the torsion fiber does not twist by matching the rotating rate of the two turntables. As a result, this method is insensitive to the fiber's properties. In this work, several sources of uncertainty existed in our preliminary experiment [4] are reduced: (1) An ultra-low thermal expansion (ULE) glass shelfis used to support the source masses instead of the aluminium shelf to reduce the temperature influence on the distance between source masses; (2) The background gravity gradient of the rotating shelf is measured and compensated with small blocks so that its effect to the $G$ value is reduced to ~1 ppm; (3) Two different methods are used to check the source mass distance to improve the confidence level. With the improvements shown above and a lot of careful systematic error evaluation, a new $G$ value with a prospective uncertainty of less than 15 ppm will be obtained.

Increasingly sophisticated code obfuscation techniques are quickly adopted by malware developers to escape from malware detection and to thwart the reverse engineering effort of security analysts. State-of-the-art de-obfuscation approaches rely on dynamic analysis, but face the challenge of low code coverage as not all software execution paths and behavior will be exposed at specific profiling runs. As a result, these approaches often fail to discover hidden malicious patterns. This paper introduces SEEAD, a novel and generic semantic-based de-obfuscation system. When building SEEAD, we try to rely on as few assumptions about the structure of the obfuscation tool as possible, so that the system can keep pace with the fast evolving code obfuscation techniques. To increase the code coverage, SEEAD dynamically directs the target program to execute different paths across different runs. This dynamic profiling scheme is rife with taint and control dependence analysis to reduce the search overhead, and a carefully designed protection scheme to bring the program to an error free status should any error happens during dynamic profile runs. As a result, the increased code coverage enables us to uncover hidden malicious behaviors that are not detected by traditional dynamic analysis based de-obfuscation approaches. We evaluate SEEAD on a range of benign and malicious obfuscated programs. Our experimental results show that SEEAD is able to successfully recover the original logic from obfuscated binaries.

Big data analytics frameworks are developing towards larger degrees of parallelism and shorter task durations to achieve lower latency. Consequently, millions of scheduling decisions need to be made per second, which has posed a big challenge to today's centralized schedulers. Therefore, many researchers and enterprises turn to distributed scheduling approaches to avoid the throughput limitation of centralized designs. To our knowledge, Omega, Apollo and Sparrow are three famous approaches that make prior moves in distributed scheduling but they each have shortcomings and none of them try peer-to-peer architecture. We then propose a new scheduling approach called Piper that adapts peer-to-peer idea to the domain of distributed scheduling, which provides near-optimal performance. We have implemented Piper using Apache Thrift and the results show that Piper reduces job response times by over 1.5× when compared to Sparrow (we select Sparrow for comparison because it is a leading design and has been open source). In addition, trace-driven simulations have been used to evaluate Piper when scaling to large clusters, which further reveals that Piper provides better performance than Sparrow.

Gene expression programming (GEP) is a new evolutionary algorithm, which has the very good applications in the field of function finding. In view of the insufficiency of traditional GEP, this paper puts forward an improved gene expression programming algorithm based on hybrid strategy (HSI-GEP). This paper has two improvements: (1) using mirror and reset mechanism to replace the inferior individuals of population, to improve the quality and the diversity of population; (2) introducing the clonal selection before tournament selection in order to improve the mining ability of algorithm about the superior individuals. The experiments compared with the improved GEP from authoritative literatures about function finding problems have been carried on, and the results show that HSI-GEP is of high quality, has fast convergence rate and obvious competitiveness.

The development of a city gradually fosters different functional regions, and between these regions there exists different social information due to human activities. In this paper, a Region Activation Entropy Model (RAEM) is proposed to discover the social relations hidden between the regions. Specifically we segment a city into coherent regions according the base station (BS) position and detect the stay and passing regions in trajectories of mobile phone users. We regard one user's trajectory as a short document and take the stay regions in the trajectory as words, so that we can use Natural Language Processing (NLP) method to discover the relations between regions. Furthermore, the Region Activation Force (RAF) is defined to measure the intensity of relationship between regions. By measuring the Region Activation Entropy (RAE) based on RAF, we find an 88% potential predictability in regional mobility. The result generated by RAEM can benefit a variety of applications, including city planning, location choosing for a business and predicting the spread of human. We evaluated our method using a one-month-long record collected by mobile phone carriers. We believe our findings offer a new perspective on research of human mobility.

Inspired by the evolution process of human, a double system co-evolutionary gene expression programming (DSCE-GEP) is proposed in this paper. The DSCE-GEP consists of natural evolution system and artificial intervention system, where the natural evolution system solves problem according to the mode of standard gene expression programming (ST-GEP) and the artificial intervention system guides the process of natural evolution system to solve problem. The artificial intervention system includes individual intervention operation and population intervention operation and is based on a gene pool, which is first created according to prior knowledge and then is updated according to posterior knowledge, contains eminent genes, morbid genes and their characteristic information as well. The individual intervention operation, which aims to improve the quality of population individuals, consists of a repairing operator that removes the morbid genes in individuals and a strengthening operator that spreads eminent genes to the individuals of population. The population intervention operation uses entropy as the diversity indicators of population and increases the population entropy by introducing a certain number of feasible random individuals and feasible mirror individuals to improve the global searching ability of the algorithm when the population entropy drops to a given threshold. The experimental results show that the performance of DSCE-GEP is better than other GEP algorithms proposed in the related literatures as regards function finding problems, and promises competitive performance.

There are huge amounts of User Generated Contents (UGCs) consisting of authors' articles of different themes and readers' comments in social networks every day. Generally, an article often gives rise to thousands of readers' comments, which are related to specific points of the originally published article or previous comments. Hence it has suggested the urgent need for automated methods to implement the content linking task, which can also help other related applications, such as information retrieval, summarization and content management. So far content linking is still a relatively new issue. Because of the unsatisfactory of traditional ways based on feature extraction, we look forward to using deeper textual semantic analysis. The Word Embedding model based on deep learning has performed well in Natural Language Processing (NLP), especially in mining deep semantic information recently. Therefore, we study further on the Word Embedding model trained by different neural network models from which we can learn the structure, principles and training ways of the neural network based language models in more depth to complete deep semantic feature extraction. With the aid of the semantic features, we expect to put forward a new method for content linking between comments and their original articles in social networks, and finally verify the validity of the proposed method through experiments and comparison with traditional ways based on feature extraction.

Transit Signal Priority (TSP) is a computer network control system. However, it is inaccurate to input traffic flow by the method of random distribution and analyze the delay. With communication technology developing precisely calculating the delay time of buses can be achieved. The kernel algorithm of TSP still has room for improvement. In this paper, the change rate of movement track of bus is studied under the condition of new communication technology. According to the characteristics of real-time monitoring data, the calculation model of average person delay was proposed by utilizing exponential smoothing method. A case study was conducted to validate the model results. Simulation results shows that this method effectively reduces average delay time of the travelers 20.9%.

In the future of display technology, laser display with its realistic display effect has drawn attention of people all over the world. In order to make the exiting color signal and laser display compatible, the existing color signal must be converted into laser color gamut. This paper according to the trait of laser display, makes the Xilinx System Generator as the foundation, uses the modularized way to realize the gamut conversion from the existing display color gamut to the laser color gamut. System Generator generated IP Core which is easy to transplantation and modification and getting software and hardware collaborative simulation results through the Xilinx FPGA development platform.

Traffic monitoring is not only used to collect traffic information but also used to analyze the cause of traffic accident. However, the accuracy of video image detection system is dependent upon many factors such as the camera height, location, and angle above the roadway. There is small number of research in this field at present. This paper discusses the effect of different camera angles on the precision of parameters for analysis of traffic accident. The angle scope ranges from 0 degree to 90 degrees. The result shows that camera angle plays a key role in the precision of parameters for traffic accident analysis. A corresponding relationship existed deeply between the precision of parameters and camera angle. The analytical results lead to recommendations on use of video to detect traffic parameters.

Objective: To observe the feasibility of virtual organ computer-aided analysis (VOCAL) technique in estimating the fetal lung volume, and to assess the diagnostic value of ultrasound fetal lung volume in fetal pulmonary hypoplasia (PH). Methods: A total of 31 fetuses (20-32 weeks) with abnomal diagnosed with prenatal ultrasonography were enrolled in this study. The autopsy and clnical diagnosis were permitted. The fetal lung volume was estimated with four-dimensional ultrasonography using VOCAL technique, the value of fetal lung volume in the diagnosis of fetal PH was evaluated. Results: The sensitivity value of fetal lung volume in diagnosing fetal PH was 81.1% and the accuracy ratio was 76.92%. Conclusion: The fetal lung volume can be estimated precisely with four-dimensional ultrasonography using VOCAL technique. Fetal lung volume may be useful in the prenatal diagnosis of fetal PH.

In order to research on the water filling and emptying process of a pressured-pipeline in long distance, according to the characteristics for the pressured-pipeline in the long distance during the filling and emptying process, the strategies of sectional treatment and water filling and emptying in succession were adopted. For each segment of the pipe, a mathematical model was established with using the ideal gas equation according to the filling principle under the low flow rate in combination with the characteristics of inlet and exhaust gas of an air valve. On the basis of the definite positions of air valves and drain pipes, calculating the air pressure changes in different pipes according to the definite water filling discharge and different water emptying discharges. The calculated results indicate that in accordance with the principles of the low flow rate of water filling, the air pressure inside the pipe is not high. The greater the water emptying discharge, the lower the corresponding air pressure and the higher pipe risk. These computation results can provide safety assessment foundations for the pressured-pipeline filling and emptying process in the long distance.

A systematic pot test was performed both in summer and autumn, which the purification efficiency to the polluted urban storm runoff was studied. The orthogonal method was applied to design the experiment. 24 plastic pots were divided into eight groups, and each group included a blank sample and two parallel samples for comparison. The results showed that: (1) the removal rate of COD Cr was almost higher than 90%, which was mainly dependent on the absorption of substrate. And the positive correlation between removal rate and pollutant concentration was existed; (2) the removal rate of TN in plant pots was obviously better than that of no plant; (3) the removal rate of TP was assured by the absorption and complexation of substrate, which was higher than 90%; (4) the experiment was performed on the condition of intermittent inflow and the higher removal rate was reached, which indicated that the purification function of urban riverbank could recover in the period without storm.

The time varying characteristic of pressure fluctuation was analyzed with the physical model test of a new pumping station in South-to-North water diversion. In order to study the non-stationary property of pressure fluctuation signal, the empirical mode decomposition (EMD) method was used to perform time and frequency analysis. And the correlation Coefficients of intrinsic mode function were calculated to identify the frequency components and their time variation with the operation conditions. The analysis results showed that EMD method can improve the resolution of time frequency analysis and extract main vibration components from vibration signal more clearly.

In this paper, we propose a distributed call admission control scheme for multiservice OFDMA system. In our proposed call admission control (CAC) scheme, a cell dynamically adjusts optimal new call and handoff acceptance ratios by three steps. First, exchange numbers of calls in handoff area with its adjacent cells. Then, dynamically allocate resource for different types of traffic. Finally, use a Markov queue model and a bidirectional iterative search method (BIS) to calculate the new call and handoff acceptance ratios to obtain a near-maximal network utility which depends on both packets delay and the number of ongoing calls. Simulation results show the performance gain of the proposed scheme.

On the basis of analyzing the insufficiency of genetic algorithm in the solution of distribution network reconfiguration, an improved immune genetic algorithm (IIGA) is proposed. The key of IIGA lies in the construction of vaccine pool and the design of immune operator. The vaccine pool can be created and updated automatically, and the immune operator consists of vaccination and immunoassay. In addition, the encoding method based on the fundamental loop, hyper-mutation and tournament selection operator with big selection pressure parent-offspring competition are adopted. Simulation results of IEEE 33-bus system and IEEE 69-bus system show that IIGA conforms to the feature of distribution network reconfiguration and can effectively restrain both degeneration and fluctuation phenomena during the evolution; it possesses fast convergence speed while the quantity of the solution is ensured. Comparing with traditional genetic algorithm (CGA) and allied algorithms in related literatures, IIGA promises excellent performance not only in convergence speed but also in quality of solution.

In this paper, we present the design of an asynchronous 8051 microcontroller with interface to external commercial memory. The design consists of an asynchronous core implemented using dual-rail four-phase protocol, a 128 byte internal asynchronous RAM and other synchronous peripherals including interrupts, timers and serial port. The asynchronous core contains all standard 8051 instructions except for multiplication and division. The interface to external commercial ROM and SRAM is controlled by two internal counters with adjustable overflow values to accommodate potentially variable clock source and external memory access time. An acknowledge signal is generated once the counter overflows which indicates the completion of a read/write operation. The chip is implemented using Austria Micro Systems (AMS) 0.35µm technology. It is able to operate at 0.22MIPS and consume 141pJ/Instruction at 1.0V supply. Another two-stage pipelined version designed later operates at 0.3MIPS and consumes 180pJ/Instruction at the same supply1.

On the basis of the analysis of canonical genetic algorithm's shortage, a novel genetic algorithm based on dynastic changes mechanism of nation (DCGA) is proposed. In DCGA, populations one by one evolve according to the relay race mode, by which the DCGA holds a mechanism of transferring from a local optimal solution to another better one and finally finding the global optimal solution under the guidance of knowledge, which is continuously updated along with the alternation of population evolution. To evaluate DCGA, TSP is applied. The results of simulation show that DCGA can avoid premature convergence and evolutionary stagnation efficiently and has the strong ability of realizing the global optimization.

In this work, a novel approach using probabilistic CMOS (PCOMS) technology is used to reduce the energy consumption of a fuzzy PID (proportional-integral-derivative) controller. Energy saving is achieved through designing a probabilistic circuit which deliberately reduces the supply voltage of some less significant bits. The fuzzy PID controller consists of 15 bits with floating point representation. Through numerical simulations and VHDL validation, the fuzzy PID can obtain a satisfactory tradeoff with about 4% deviation while achieving a total energy saving of about 32%. Through error analysis, a fuzzy PID is redesigned to tolerate more of randomness in the control signals, hence obtain a better steady state performance while achieving an energy saving of about 51%.