Entity

Time filter

Source Type


Zhang X.,PLA Second Artillery Engineering University | Wang H.,PLA Second Artillery Engineering University
Yi Qi Yi Biao Xue Bao/Chinese Journal of Scientific Instrument | Year: 2010

The threshold is the key factor of threshold-based wavelet denoising. To achieve the optimal results of denoising signals corrupted by Gaussian noise, a simplified cross-validation (SCV) algorithm was proposed to estimate the optimal threshold. In this algorithm, due to the similarity between the odd-indexed parts and even-indexed parts of the signal, the computation cost of cross-validation (CV) algorithm was reduced, and particle swarm optimization (PSO) was applied to realize the evolutionary search of the optimal threshold. Comparative experiments on simulated signals and rotor fault signals were carried out to confirm the effectiveness of SCV. The experimental results indicated that SCV improves the computation efficiency of CV and outperforms the conventional threshold-based wavelet denoising methods in noise reduction. Source


Zhang X.,PLA Second Artillery Engineering University | Wang H.-L.,PLA Second Artillery Engineering University
Wuli Xuebao/Acta Physica Sinica | Year: 2011

To solve the problem of extreme learning machine (ELM) on-line training with sequential training samples, a new algorithm called selective forgetting extreme learning machine (SF-ELM) is proposed and applied to chaotic time series prediction. The SF-ELM adopts the latest training sample and weights the old training samples iteratively to insure that the influence of the old training samples is weakened. The output weight of the SF-ELM is determined recursively during on-line training procedure according to its generalization performance. Numerical experiments on chaotic time series on-line prediction indicate that the SF-ELM is an effective on-line training version of ELM. In comparison with on-line sequential extreme learning machine, the SF-ELM has better performance in the sense of computational cost and prediction accuracy. © Chinese Physical Society. Source


Han B.,PLA Second Artillery Engineering University | Wang Y.-M.,PLA Second Artillery Engineering University
Binggong Xuebao/Acta Armamentarii | Year: 2010

In grayscale image matching, the normalized cross correlation algorithm has a lot of advantages such as robust, high precision, fitting for implementing on hardware and so on. But its application is limited because of high algorithm complication and low computation speed. Therefore, a fast normalized cross correlation algorithm was proposed based on iterative idea. At computing the energy of sub-based image, the sum of all the pixels in sub-based image is fastly computed by adding and subtracting several pixels' values to reduce the algorithm complication. The experimented results show that, the fast normalized cross correlation algorithm has the same matching precision as that of the original one, but its implementation time is just 1/9 of that of the original. Source


Li B.-P.,PLA Second Artillery Engineering University
Baozha Yu Chongji/Explosion and Shock Waves | Year: 2010

Guided bombs were taken as the research objects to numerically simulate the penetration and explosion process of large-aperture-heavy weapons by using a three-dimensional arbitrary Lagrange-Eulerian (ALE) code and to explore damage effects of a roller compacted concrete gravity dam under continuous attacks of two precision-guided bombs. Results show that the vibration caused by penetration is weak, while the particle vibration is mainly caused by ammunition explosion. The intermission time of sequential attacks is far longer than the dynamic response stretch of the dam, and the dam vibration caused by sequential attacks scarcely takes on the superposition effect. But the damage effect of penetration and explosion for the former missile provides the freedom surface for the subsequent missile, and the penetration depth and explosion-damaged range of the subsequent missile increase. The explosion-damaged ranges basically connect with each other, which threat to the normal running and security of the dam. Source


Zhang X.,PLA Second Artillery Engineering University | Wang H.-L.,PLA Second Artillery Engineering University
Kongzhi yu Juece/Control and Decision | Year: 2012

To solve the problem of extreme learning machine(ELM) on-line training, an algorithm, fixed-memory extreme learning machine(FM-ELM), is proposed. FM-ELM adopts the latest training sample and abandons the oldest training sample iteratively to enhance its adaptive capacity. The output weights of FM-ELM are determined recursively based on Sherman-Morrison formula. Thus, the computational cost of FM-ELM training procedure is effectively reduced. Numerical experiments on nonlinear system on-line condition prediction show that FM-ELM has better performance in adjusting speed and prediction accuracy in comparison with on-line sequential extreme learning machine(OS-ELM). Source

Discover hidden collaborations