Applied Mechanics and Materials Vols. 556-562

Abstract: Aiming at the problem that flight data recorder taken aircraft directly as testing signal resource in the process of development and debugging of the flight data recorder is very difficult, the flight data signal generator based on embedded system is designed, which purpose is to simulate the status signal of aircraft and each component of aircraft during flight, providing test signal source for flight data recorder. Interactive interface was designed on PC by LabVIEW, which was used for setting parameter values of each signal; STM32 microcontroller received the parameter value and controlled each generating module to generate electric signal. All generating modules link with the Bus network, and the network will be easy to increase or decrease generating module. Experiment results show that the flight data signal generating system can simultaneously generate 138 signals, including analog, digital, synchronous, frequency, ARINC429 bus signals etc, and can work stably.

Abstract: The traditional graph cut algorithm for complex image matting accuracy is not high, the weight parameters in the dynamic graph cut energy model is based on image matting. The degree of similarity the algorithm uses Gauss model to calculate the foreground color and background color, in graph cut energy model of dynamic weight parameters are added, by it and then many corrections to the image matting. The experimental results show that the algorithm is easy to operate, that can improve the accuracy of image matting.

Abstract: In a frequency selective fading environment amplify and forward relay system, the channel estimation algorithm and least squares (LS) estimation of mean square error is relatively high. And when using the compressed sensing (CS) which was proposed recently in orthogonal matching pursuit algorithm (OMP) for the single relay channel and multi relay system estimation, in order to reduce the mean square error of the system. In multi-relay cooperative communication system, allocated the power to the relay node, then respectively using the LS and OMP for channel estimation. The results show that, the orthogonal matching pursuit algorithm can greatly reduce the system mean square error, and the same channel estimation algorithm was used for the relay power allocation is also effectively reduce the estimation error, to further improve the accuracy of channel estimation.

Abstract: Time delay estimation is a key technology of array signal processing, which is designed to estimate the time delay of the sound source signal reaches different sensor. Because of differences in the weighting function method, existing methods of generalized cross-correlation time delay estimation have their own anti-noise and anti-reverberation characteristics. In this paper, the existing various delay estimation algorithm comprehensive is compared, which reveals their advantages and disadvantages, and further improved direction of time delay estimation algorithm is given.

Abstract: InGaAs is direct and narrow bandgap material with ultrahigh electron mobility, and is a promising candidate for optoelectronic device in the near-infrared region. The main objective of this manuscript is to design an InGaAs semiconductor-based photodetectors nanowire infrared photodetector, which would be manipulated in optical response wavelength range at room temperature with high-detective and fast-responsive performances. Considering into account mature technology in laser device and for maximizing 1.55 um optical communication performance, the design of achieved bandwidth is >1 Gbps. According to the theoretical calculations, the fundamental parameters of the InGaAs core-shell nanowire APD photodetector device are obtained, with certain values of single nanowire diameter, lateral size and applied voltage, r=100 nm, l=1 μm, V_bias=10 V. Meanwhile, we deal with the electrons mobility of the internal nanowire based on three factors, size effect, temperature and electric field effect. The results demonstrate that the InGaAs nanowire APD is potential candidatefor high rate in satellite laser communication field.

Abstract: The author proposes a novel DCT domain image encryption algorithm based on Lorenz chaotic system. The paper firstly subdivides digital image into sub-images, each of which is transformed with DCT transform. Then some DCT transform coefficients are selected with filtering. Finally the retained DCT transform coefficients are cross encrypted with three chaotic sequences which Lorenz chaotic system generates. Simulation experiments and secure performance analysis show that the algorithm has a good encryption effect and anti-attack abilities.

Abstract: Based on the fabric of national electric power system, the types and causes of delay in the network is investigated. The evolution guideline is deeply discussed. Mechanism in T-MPLS such as packet switching, plastic limit, automatic protection switching and multi-service package adapter are utilized to improve the time delay of the power communication network. It is to lay the foundation for the next generation of Smart Substation.

Abstract: In this paper, we have designed two experiments to investigate the visual comfort of focus pulling for S3D video with different shooting configurations. The variables mainly includes:1) convergence setting ;2) the distance between camera and subjects;3) pulling direction of focus. Evaluation results showed that different convergence settings would affect the visual comfort during focus pulling obviously. With the increasing of convergence distance, the subjects would appear from inside to outside of the screen, and the visual comfort value of focus shifting would decrease. Our research can provide data analysis for improving the visual comfort and shooting technology of S3D videos.

Abstract: The Context-Dependent Deep-Neural-Network HMM, or CD-DNN-HMM, is a powerful acoustic modeling technique for HMM-based speech recognition systems. The CD-DNN-HMM can greatly outperform against the conventional Gaussian-mixture HMMs. Therefore, we build a CD-DNN-HMM LVCSR system by modifying a mature GMM-HMM system. The baseline CD-DNN-HMM system achieve word-error rate of 18.6% that is far better than 24.9% achieved by the GMM-HMM system. However, the speed of the baseline CD-DNN-HMM system becomes a major roadblock for its real-time rate reaches 0.72 on the standard NIST 2000 Hub5 evaluation set. In this paper, we realize several optimization algorithms in our baseline system to accelerate the recognition speed. Testing the optimized system on the same evaluation set, we achieve real-time rate of 0.39, a relative reduction of 45.8%.

Abstract: Memory forensics can provide crucial evidence or further investigation clues, it also has a direct connection with field investigation. Image is a common and intuitive way to express information, and has direct effect as evidence. This article analyzes the characteristics of common image formats (BMP, PNG, JPEG, GIF) data stored in the memory and summarizes some rules of extracting the picture data. On this basis, the article proposed a method of extracting picture data from the memory image. The experimental results show that picture data can be accurately recovered from the image.