Search

In the context of the current large number of malicious codes, the detection and protection of malicious codes is particularly important. In recent years, a method of using deep learning to detect malicious code has emerged. Thus, in this paper, we propose a new detection method that converts binary files of malicious code into decimal arrays and use 1-D CNN to perform classification and recognition. Aiming at the imbalance in the number of code families, we choose xgboost, which performs well in the classification prediction competition. We conduct experiments on 9,458 malware samples from 25 different malware families in the Vision Research Lab. The experimental results show that our classification prediction reaches 97% accuracy.

Hollow core slabs are widely used in prefabricated buildings either as slab elements or bridge decks because of their economic benefits. However, little attention has been given to developing strengthening techniques that are especially suitable for hollow core slabs to recover or improve the original serviceability and load carrying capacity. In this study, two novel techniques for strengthening hollow core slabs with mounted steel bars and prestressed steel wire ropes (P-SWRs) were proposed. Six slab specimens were tested to failure under a four-point bending configuration, and the strengthening effects of the mounted steel bars and P-SWRs on the hollow core slab performance were studied. The experimental results indicated that the steel bars mounted in the cores were capable of enhancing the flexural performance of the hollow slabs. The prestressing technique and anchorages developed in this paper could be used to feasibly and easily prestress the steel wire ropes. P-SWRs were found to be more efficient in improving cracking load than mounted steel bars because of the prestress in P-SWRs. External P-SWRs were better at improving the cracking load, yield load and precracking stiffness compared with internal P-SWRs. The simultaneous use of both internal and external P-SWRs obtained the highest flexural capacity. A three-dimensional finite element (FE) model was developed incorporating the smeared-crack model for concrete and elastic-linear strain hardening model for steel. The validity of the FE model was verified through comparisons with experimental results. Thus, this study provides both experimental evidence and finite element model for the application of the proposed strengthening techniques for hollow core slabs.

Band selection is widely used to identify relevant bands for land-cover classification of hyperspectral images. The combination of spectral and spatial information can improve the classification pe...

Due to the controllable bond performance and high-efficiency construction of bundled hybrid reinforcements, the seismic performance and accelerated construction of a precast concrete structure can be guaranteed. Eighteen pullout specimens of bundled deformed steel bars and basalt fiber reinforced polymer (BFRP) bars (single, double, three-bundle) were tested under monotonic loading. The peak bond strength decreased by 40.7% as the reinforcement bundles increased from one to three. The corresponding experimental investigations on five concrete beams reinforced by different types of bundled reinforcement showed that, compared to a beam with ordinary distributed reinforcement (control beam), the ultimate displacement of the beams with three bundles and six bundles increased by 60% and 90%, respectively. The load bearing capacity of the six-bundle beam was 9.6% larger than that of the control beam. A simplified bond-slip model based on the concept of equivalent bond area was proposed and showed good agreement with the test results. In addition, finite element analysis considering the bond-slip between the reinforcement and concrete was conducted to predict the load–deflection relationships of the new composite beam. A comparison of the analytical and experimental results showed that the analytical model could accurately predict the failure modes, load–deflection relationships and strain developments of bundled hybrid reinforced beams.

Hyperspectral sensors often collect hundreds of bands at a time, so hyperspectral images can accurately characterize different land-cover types with abundant spectral information. However, these spectral bands also contain redundant information that needs to be removed. Band selection is one of the most widely used methods to remove noised or redundant bands. Because labelled samples are difficult to collect, most band selection methods adopt unsupervised ways to select diverse and representative bands. Still, noised bands are often selected because they usually have low correlation with other bands. In this article, objective image quality assessment is introduced to indicate the quality of every band, and combined with the redundancy measure, a new unsupervised band selection method is proposed. Three real hyperspectral images are used to demonstrate the effectiveness of the proposed algorithm.

The eccentricity is the main concern when installing a benchmark circular grating, because it accounts for the majority of the angular measurement error. A new method to identify the magnitude and direction of the eccentricity after the installation was proposed. Compared with models in the form of Fourier series, the model of the new method has fewer coefficients and a better error compensation effect. According to the calibration experiment, putting the new model into practical error compensation, the uncertainty of the angular measurement was reduced by 79.4%.

High-precisely measuring equivalent series resistance (ESR) is necessary to improve the performance and quality of supercapacitors and enhance efficiency of energy transformation. Based on the equivalent model of supercapacitor, using the measuring method of direct and constant current discharge voltage jump, precision of measurement was raised by larger direct and constant current; synchronous sampling reduced the requirement of constant current; affectation of contact resistance was eliminated; and error resulted from ambient temperature was reduced. After comparison experiments, the results show that ESR error is less than 0.05m Ω by measuring 350F supercapacitors. The requirement of high-precision measurement can be met.

Friction is a key factor that influences the low velocity performance of high precision servo systems. It usually causes the steady-state error of servo systems and results in the creeping phenomenon in low velocity motions. Friction is complicated nonlinear process with time-varying uncertainties. The traditional PID controller of servo systems with fricition was not satisfactory in control precision and response time. In this paper, the model of servo system is obtaind through open-loop u-Ω curve and the compensating voltage for friction torque is estimated so as to eliminate the friction disturbance via feedforward compensation. A new algorithm that combines the advantages of ADRC(active disturbance rejection controller) and friction torque feedforward compensation is proposed. Comparison experiments with the new ADRC technique and the common PID algorithm are carried out on the servo system. The result demonstrates that the improved ADRC technique could reduce steady-state error to only one third of the one using the PID algorithm.

This paper develops a new real-time single object tracking system under camera rotation freely. Optical-flow and color information are combined and it proposes a simple and feasible method by updating the feature point to update the template. By capturing the corner and combining with pyramid-optical-flow the problem of real-time tracking in optical-flow is solved. The rotation and tracking under partial occlusion are settled through introducing the color probability distribution and recapturing the feature points. It is proved by experiment that this system is insensitive to the environment and shows a good performance.

In current version of Trusted Network Connection (TNC) architecture, IF-MAP protocol creates a structured way to support real-time data interchange and dynamic management. However, there is a big challenge that the MAP Server has to process mass metadata published by all kinds of other security components in trusted network. In this paper, we propose a novel design for MAP (Metadata Access Point) Sever based on information fusion. A module called HLIF (High level information fusion) is implanted to the current MAP Server. The HLIF module correlates metadata and projects threats in the trusted network to ensure the MAP Clients can share real-time information efficiently and make precise policy decisions in anticipation. Detailed analysis reveals that HLIF module effectively enhances the MAP Server's ability of protecting security components from network attacks in trusted network.

In order to reduce the risk and cost, manufacturing enterprises turn to postponement strategy to meet differential demand of customers: this is helpful to enhance the customer service level, and this is also helpful to achieve the goal of maximizing the inventory profit. Under this assumption, inventory batch allotment shows new features, therefore, the paper proposes an optimal inventory model based on postponement strategy, which includes the framework model and the decision-making model. The framework model is firstly discussed, and then its efficiency is verified. Secondly, the decision-making model of inventory batch allotment is established with a bi-level programming approach. The decision-making model is calculated and verified using the investigation data, which is proved to be efficient to give reasonable allotment of supply chain inventory batch. It also provides an effective decision-making tool of supply chain coordination.