Search

Understanding questions expressed in natural language is a fundamental challenge studied under different applications such as question answering (QA). We explore whether recent state-of-the-art models are capable of recognizing two paraphrased questions using unsupervised learning. Firstly, we test QA models’ performance on an existing paraphrased dataset (Dev-Para). Secondly, we create a new paraphrased evaluation set (Para-SQuAD) containing multiple paraphrased question pairs from the SQuAD dataset. We describe qualitative investigations on these models and how they present paraphrased questions in continuous space. The results demonstrate that the paraphrased dataset confuses the QA models and decreases their performance. Visualizing the sentence embeddings of Para-SQuAD by the QA models suggests that all models, except BERT, struggle to recognize paraphrased questions effectively.

In terms of the motion planning problem of spacecraft proximity operations with obstacle avoidance under low uncertainty, the improved equal-collision-probability-curve and improved linear quadratic regulator (IECPC-ILQR) strategy is proposed. Firstly, the novel function of the IECPC algorithm is developed to generate the avoidance control impulse. Subsequently, the ILQR is designed to track the reference trajectory. Furthermore, combining the improved ECPC algorithm with the ILQR controller, the composite controller of the IECPC-ILQR strategy is obtained and is implemented on the chaser spacecraft. Compared with the traditional ECPC algorithm, the IECPC-ILQR strategy can avoid collision in the presence of low uncertainty. Furthermore, the proposed avoidance strategy can obtain higher control precision while requiring the same fuel. Finally, numerical simulations verify the effectiveness of the proposed IECPC-ILQR strategy.

Although the mechanisms of steady-state visual evoked potentials (SSVEPs) have been well studied, none of them have been implemented with strictly experimental conditions. Our objective was to create an ideal observer condition to exploit the features of SSVEPs. We present here an electroencephalographic (EEG) eye tracking experimental paradigm that provides biofeedback for gaze restriction during the visual stimulation. Specifically, we designed an EEG eye tracking synchronous data recording system for successful trial selection. Forty-six periodic flickers within a visual field of 11.5° were successively presented to evoke SSVEP responses, and online biofeedback based on an eye tracker was provided for gaze restriction. For eight participants, SSVEP responses in the visual field and topographic maps from full-brain EEG were plotted and analyzed. The experimental results indicated that the optimal visual flicking arrangement to boost SSVEPs should include the features of circular stimuli within a 4–6° spatial distance and increased stimulus area below the fixation point. These findings provide a basis for determining stimulus parameters for neural engineering studies, e.g. SSVEP-based brain-computer interface (BCI) designs. The proposed experimental paradigm could also provide a precise framework for future SSVEP-related studies.

Gaofen-4 (GF-4) satellite is the first high-resolution geostationary orbit (GEO) optical satellite in the world, which has appealing potential in wide-area maritime surveillance and can provide dynamic information of ships as its high revisit time. However, its spatial resolution is not high enough to obtain attributes of ships, while Gaofen-1 (GF-1) satellite in low earth orbit (LEO) can extract rich features, as its high spatial resolution. Therefore, the integration of GF-1 and GF-4 satellites can provide a better maritime situational awareness. The aim of the paper is to provide a feasible architecture of data fusion of GF-1 and GF-4 imagery for maritime surveillance, and a novel ship association method using multi-level information in order to improve the association accuracy, and experimental results verify the effectiveness of our method.

The grid impedance is widely recognized as the key factor to the stability of the grid-tied power electronic devices. However, the influence of the grid impedance ratio is always neglected due to the assumption that grid reactance is much large than grid resistance. In this paper, it is found that the grid impedance has a different impact on the stability with the reduction of grid impedance ratio. A large grid resistant can improve the stability of the grid-tied VSC due to that the grid resistance can increase the magnitude margin of the grid-tied VSC open-loop SISO transfer functions. Simulations are carried on Matlab/Simulink to verify the effectiveness of the proposed analysis.

A semantic tableau method, called an argumentation tableau, that enables the derivation of arguments, is proposed. First, the derivation of arguments for standard propositional and predicate logic is addressed. Next, an extension that enables reasoning with defeasible rules is presented. Finally, reasoning by cases using an argumentation tableau is discussed.

One method to solve expensive black-box optimization problems is to use a surrogate model that approximates the objective based on previous observed evaluations. The surrogate, which is cheaper to evaluate, is optimized instead to find an approximate solution to the original problem. In the case of discrete problems, recent research has revolved around discrete surrogate models that are specifically constructed to deal with these problems. A main motivation is that literature considers continuous methods, such as Bayesian optimization with Gaussian processes as the surrogate, to be sub-optimal (especially in higher dimensions) because they ignore the discrete structure by, e.g., rounding off real-valued solutions to integers. However, we claim that this is not true. In fact, we present empirical evidence showing that the use of continuous surrogate models displays competitive performance on a set of high-dimensional discrete benchmark problems, including a real-life application, against state-of-the-art discrete surrogate-based methods. Our experiments with different kinds of discrete decision variables and time constraints also give more insight into which algorithms work well on which type of problem.

This paper first establishes a photovoltaic generator electromechanical transient simulation model considering low voltage ride through (LVRT) and its recovery process dynamic simulation, and then studies the impact of photovoltaic low voltage ride through characteristics on the transient voltage rise of new energy generator under large disturbance impact. Based on the analysis and summary of the model structure and corresponding mechanism of distributed new energy in China, the influence of photovoltaic low voltage ride through characteristics on the operation characteristics of receivingend power grid is studied.

On July 2nd, 2017, Pole I of Binjin UHVDC failed to recover after two restart attempts, which impact both the sending-end and receiving-end system. First of all, the detailed process of restart function is presented and the impacts on frequency and voltage are analyzed. Secondly, the main impacts on both sending-end and receiving-end system are studied by replaying the failed recovery of Binjin UHVDC in the simulation, and the possible risks caused by improper restart strategy, unreasonable power flow arrangement and malfunctions of HVDC control system in the process are analyzed. On this basis, some suggestions are proposed in order to reduce the impact of HVDC restart and ensure security and stability of both sending-end and receiving-end system.

Big data is one of the keys to unlock the next great waves of economic growth. A number of smart-world systems (smart grid, smart transportation, smart cities, etc.) driven by advances in Internet of Things (IoT) technologies will accelerate the generation and collection of big data to an unprecedented scale. As a fundamental commodity in our current information age, big data is the crucial competitiveness in modern commerce. This calls for the development of a big data market: pricing, trading, privacy preserving and protection, data security, etc. In this talk, we will discuss the related issues of big data market: pricing, trading, more specifically the big data auction, how we can preserve privacy by leveraging the concept of homomorphic cryptography and secure network protocol design. We propose a generic Privacy-Preserving Auction Scheme (PPAS), in which the two independent entities of Auctioneer and Intermediate Platform comprise an untrusted third-party trading platform. Yet, to further improve the security of the privacy-preserving auction, we also propose an Enhanced Privacy-Preserving Auction Scheme (EPPAS) that leverages an additional signature verication mechanism. The feasibilities of both schemes are validated through detailed theoretical analyses and extensive performance evaluations, including assessment of the resilience to attacks. In addition, we will discuss some open issues and possible extensions relevant to our schemes.

The purpose of this paper is to present a design of a foldable pneumatic soft manipulator. The soft manipulator is made up of three foldable pneumatic soft modules, which are stacked in series. The foldable pneumatic soft module is composed with a pneumatic actuator and an inflatable beam. The function of the pneumatic actuator is to make the directional movement, while the inflatable beam is the section with foldability. In order to get the performance of this kind of design, one of the soft modules is developed. Moreover, the pulling force of the pneumatic actuator, the bending stiffness of the inflatable beam and the pointing movement of the soft module are tested in the experiment. The concept of the foldability and extension of the soft manipulator is also validated.

A new computational model for brittle fracture and fragmentation has been developed based on finite element analysis of non-linear elasticity equations. The proposed model propagates the cracks by splitting the mesh nodes alongside the most over-strained edges based on the principal direction of strain tensor. To prevent elements from overlapping and folding under large deformations, robust geometrical constraints using the method of Lagrange multipliers have been incorporated. The model has been applied to 2D simulations of the formation and propagation of cracks in brittle materials, and the fracture and fragmentation of stretched and compressed materials.

Under the case of ill-conditioning system, this paper is concerned with the AC/DC power flow calculation. The improved Levenberg–Marquardt (ILM) method with adaptive damping factor selection is applied to solve the AC/DC power flow problem. The main purpose of this paper is as follows: one is to provide comparison reference between Newton method, classical LM method (CLM) and ILM method under the well-conditioning system; the other is to research what is the maximal load withstood by power system, under the case of ill-conditioning. Finally, those methods are tested on the 22-bus, the IEEE 118-bus AC/DC system, respectively. Numerical results indicate that the ILM method has the advantage of fast convergent speed. When expanding loads in a certain extent, ILM method can at least find least square solutions, whereas Newton method and CLM method would divergent, and the convergent property of Newton method can be improved by taking some measurements using the information of a least square solution obtained by ILM method.

Large transient power impact of UHVDC commutation failure has great influence on the security and stability of the power grid, becoming one of the most important threats to the safe operation of large scale AC and DC interconnected system, especially when the HVDC system is connected to the weak AC power grid, commutation failure can even cause system instability, leading to new energy tripping, system crash and other serious risks. A coordinated recovery strategy of AC and UHVDC interconnected system is proposed in this paper. Firstly, the coupling characteristics of AC and HVDC system is analyzed, and then the impact of different disturbance on the power system recovery characteristics is taking into consideration. Secondly, an AC and HVDC system coordinated recovery strategy is proposed considering the AC power grid strength, by modifying the parameters of CFPREV and VDCOL to control the recovery speed of HVDC system after commutation failure, which can realize fast recovery of the interconnected system and reduce the impact of HVDC commutation failure on the whole system. At last, the effectiveness of this strategy is verified by the simulation of an actual large scale AC and UHVDC interconnected power grid.

Pulsars autonomous navigation is characterized by its credible performance and high navigation accuracy, which will have broad prospects for use in deep-space autonomous navigation. In our study, we find the observation time required for different pulsars to obtain the same time of arrival (TOA) accuracy is different, which means the rates of different sensors are different. However, the previous studies assume the pulsars used for navigation have the same observation time. And it will greatly influence navigational accuracy. Therefore, a new multi-rate observation model is presented and multi sensor data fusion is applied to the X-ray pulsar navigation. Combined with unscented Kalman filter (UKF) through simulation, this multi-rate observation model can improve the filtering accuracy effectively.

Concerning the problem that the existing direct anonymous authentication scheme cannot achieve effectively in different domains,based on the original direct anonymous authentication scheme,in this paper a direct anonymous authentication scheme was designed.It took the certificate issuer outside the domain as a proxy,and the certificate was issued by certificate issuer outside the domain directly,which solved the privacy protection problem of trusted computing platform effectively in different trusted domains,and it had permission settings to trusted computing platform in different trust domains.Demonstrated by analysis,this new scheme meets the requirements on anonymity,unforgeability and prevention of replay attacks.Furthermore,it improves the efficiency of certification scheme.

In order to improve the real-time of global motion estimation, we proposed a adaptive global motion estimation method based on improved SUSAN algorithm and SIFT algorithm. According to the five matching results before the current matching, the method uses Kalman filter algorithm to predict overlap regions of the current matching two images, and then extracts feature points in the overlapping regions instead of the whole regions of the images. In the part of extracting feature points of the method, improving the SUSAN algorithm according to the geometric characteristics of the SUSAN templates, which improves the speed of extracting feature points. Writing code to implement the method in VS2008, and experimental verification: This method accelerates the executing speed of the algorithm ensuring the accuracy at the same time.

Data traffic keeps increasing dramatically in the connected world and it pushes hard on evolution of IO technology. As technology progress, a simple and high-speed unified interconnect becomes practical to support many different protocols. Light Peak (LPK) is a high-bandwidth, inexpensive optical interconnect technology developed by Intel for this purpose. Also, to meet versatile and ever changing I/O needs, an adaptive IOH (I/O Hub) architecture could provide satisfying flexibility. In this paper, we present a high bandwidth and flexible IOH architecture for embedded system that is implemented on FPGA. The platform has 2x 10Gbps optical LPK links, and several SATA, Ethernet and Display Port interfaces. In addition to hardware, we present a software stack framework that provides a unified interface for protocols mapping, replacing numerous individual hardware controller, by utilizing LPK's versatile host controller. Thus it reduces the cost and development effort, and enhances the IOH flexibility. We also describe a configurable hardware protocol adapter in device side, which can be directly connected to any device. IO meter and Iperf show performance of SATA, Ethernet on this FPGA based IOH is competitive with the native ones.

With the development of embedded system, NAND flash memory had become a typical storage medium in embedded system. Recent years, there comes a lot of file systems special for flash memory. YAFFS is specially designed for NAND flash. YAFFS use spare area in every page in NAND flash to maintain the file structure, therefore when mount a YAFFS file system should scan entire flash memory device to construct the file structure for every single file. In this paper, we propose a new method by simulating a super block which record space usage of flash saved in a special page to overcome the problem. Test result shows that this method can reduce the mount time by maximum of 30.2% than existing method of YAFFS file system.

This paper proposed a joint adaptive soft frequency reuse and virtual cell power control (JAP) algorithm for relay enhanced cellular (REC) system in LTE-A. The JAP algorithm allocates bandwidth to base station and relay node adaptively according to user distribution and target data rate and achieves frequency reuse factor be one among sectors. Furthermore, the JAP algorithm controls transmit power based on SIR-balance principle and adjusts power in "virtual cell", which can greatly mitigate co-channel interference among sectors. Simulation results show that the proposed algorithm can obtain both high average spectral efficiency and edge user throughput.