Papers by Keyword: Singular Value Decomposition (SVD)

Abstract: While bearing fault signals are strongly interferenced by noise, diagnosis using EMD directly for bearings fault becomes incorrect. A scheme based on Singular Value Decomposition(SVD) and Empirical Mode Decomposition(EMD) is proposed for solving this problem. Aiming at bearing fault signal characteristics, SVD preprocesses sampled signals to denoise. Then preprocessed signals are analyzed by EMD. Fault characteristic frequency can be obtained by spectrum analysis for Intrinsic Mode Functions(IMFs). This method is useful to detect fault of bearings and a comparison is made between it and EMD. The results show that this scheme can diagnose fault correctly under strong noise.

Abstract: In this paper, we propose a new combined model for image de-noising. The singular value decomposition has been widely used in the image application such as de-noising till now. There is a problem, which is how to confirm the number of the singular values in the process of image reconstruction, need to solve in the process. Then the paper discusses a minimum energy model to confirm the number through the way of calculating the local minimum in the defined energy curve. The experiment results show that the established model is consistent to the fact under the situation that if the image be with regular structure.

Abstract: Background noise interference is one of the most important factors for low-voltage power line communication’s reliability. By analyzing the background noise of low-voltage power line communication’s channel, the background noise’s measuring circuit is set up and the AR model of the measured background noise is established. Both of them are respectively using singular value decomposition and Levinson-Durbin (LD) recursive method to calculate the AR model’s parameters and a comparative analysis of the simulation is made. The results induct: parameters acquired from the methods of singular value decomposition and LD recursive method are feasible, the parameter model from singular value decomposition is relatively complex, but extremely accurate, which is suitable for the off-line calculation and analysis of the low-voltage power line’s background noise; the parameter model from LD recursive method is very simple, but has a greater loss of accuracy, fitting for online quickly generation of the low-voltage power line’s background noise.

Abstract: The hidden unit number of RBF neural networks directly influences the performances of the whole net. A new strategy to prune the hidden units based on the singular value decomposition (SVD) of matrixes is proposed in the paper. At the basis of a structure involving enough more hidden units, the paper analyzes the outputs corresponding to some training samples with the SVD method and finds out the internal relations of them, then removes redundant ones according to the contribution rate of every hidden unit to the whole network, simplifies the structure of RBF neural network at last. The optimized network has strong generalization ability with simpler structure. At the end of this paper the new strategy is successfully used in the main steam system modeling of power plant and confirmed by simulation experiments.

Abstract: Recently many practical downlink multi-user MIMO linear pre-coding methods have been proposed, such as the channel inversion method and the block diagonalization method (BD). Considering the channel inversion method based on MMSE criterion (MMSE-CI) which is confined to a single receives antenna case, the BD has more advantages in multiple antennas cases, however, it has poor performance at the low and medium SNR regime on account of no consideration on the noise. In this paper, an improved MMSE pre-coding method is proposed with multi receive antennas of each user. Based on MMSE-CI, the cooperation of multiple antennas is adopted to further suppress the residual interference during designing the pre-coding matrix, which could increase the signal-to- interference-plus-noise ratio (SINR) at each user’s receiver. The proposed method obtains a better performance than the MMSE-CI and the BD algorithms, and its effectiveness is validated by both theoretical analyses and numerical simulations.

Abstract: Spatial multiplexing in the multi-user MIMO downlink allows each user in the system to receive multiple data subchannels simultaneously using the same time and spectral resources. In this paper, a successive iterative optimal algorithm based on signal-to-leakage-and-noise-ratio (SLNR) maximization algorithm is proposed, which make use of the unused subspace of some known users to improve the space gain of the other users and has no strict constraint on transmit and receive antennas numbers. According to the simulation results, the proposed algorithm outperforms the original SLNR algorithm, and has a considerable improvement in the system performance.

Abstract: With the development of computer visual technology, facial expression recognition plays an important role in the friendly and harmonious human-computer interaction field.Against the inadequacy of the original feature extraction method based on singular value decomposition, this paper proposed a hierarchical facial feature extraction method according to the needs of facial expression recognition, which combines the way of hierarchy and block to enhance the detail information of the image. Then utilize a combination of support vector machine to classify. The results of the two experiments show that the method is effective for the facial identity and expression recognition.

Abstract: Assurance of integrity of structures is the main task of structural health monitoring. The condition (health) of a structure may be revealed from its dynamic characteristics in response to natural or man-made loads. As the measurement of the natural or operating forces on actual real-life structures is generally formidable, dynamic characteristics of the structures have to be extracted from the available output signals only. The stochastic subspace identification (SSI) technique is adopted in this study to identify the equivalent system parameters of the discrete-time state equation using covariance functions of the measured output signals. With the system parameters realized, the method of damage locating vector (DLV) is then considered for further assessment. Members with nearly zero stress under the loadings of DLVs are considered potentially damaged, whereas the DLVs are derived from singular value decomposition of the change in flexibility matrix of the structure before and after the damage state. In this study, the feasibility of DLV method for damage detection of a planar structure is explored based on its seismic responses. To comply with the desired output-only scenario, the information of ground motion is discarded in the stage of SSI analysis. Despite the non-stationary nature of earthquakes, the proposed scheme has been proved sufficient for damage localization of structures from global responses (floor accelerations). The damaged locations can be identified when the structure is fully observed, regardless of single or multiple damages. In the case of partial but co-located observation (damaged floors observed), the damaged locations can still be identified with acceptable accuracy and reliability.

Abstract: Two methods of both the singular value decomposition (SVD) and the Bi-dimensional empirical mode decomposition (BEMD) were applied in extraction of gravity anomalies associated with gold mineralization in Tongshi gold field, respectively in this paper. Conclusions drawn by the comparison study are as follows: (a) The ore-controlling factor in the Tongshi gold field illustrated in the images obtained from the original gravity data by the two methods is the same that the Tongshi intrusions with a negative circular gravity anomaly and the ring contact metasomatic mineralization zone around the Tongshi intrusions with the positive gravity anomaly. (b) The two methods reveal the same spatial relationship between the ore-controlling factor and various gold mineralizations that the skarn and porphyry types of gold deposits are located within the complex pluton and the Carlin and Crypto-breccia types of gold deposits located within the contact metasomatic mineralization zone. (c) The image produced by BEMD not only reflects the structural features of the ore-controlling factor (Tongshi complex pluton), but also does the distributions of the other geological units in the Tongshi gold field such as the Mesozoic volcanic sedimentary basin in NW orientation with obvious negative gravity anomaly and the conceal metamorphic base swell in NW orientation with the positive gravity anomaly located between the Tongshi intrusions and the Mesozoic volcanic sedimentary basin. The image produced by SVD might depict in more detail the inner structure of the Tongshi intrusions and the ring contact metasomatic zone than that produced by BEMD. The higher gravity anomaly areas in island shape within the ring contact metasomatic zone may be caused by the skarn bodies with iron-copper-gold mineralization. (d) Under the constraints of the ore-forming geological setting, the results obtained from the original gravity data by combination of the two methods can depict the relationships between the ore-controlling factors and the gold mineralizations more exactly than the alternative methods.

Abstract: The combination of image features with singular value decomposition algorithm and digital watermarking algorithm based on wavelet transform respectively makes possible the advancement from pixel watermarking to content watermarking, which effectively solves the contradiction between perceptibility and robustness of conventional watermarking algorithm. Therefore, stronger robustness and lower perceptibility of watermarking are achieved. This paper first gives a review on conventional singular-value-decomposition-based digital watermarking algorithm, and then makes a thorough analysis of its respective features, advantages and defects. Improved feature extraction schemes of digital watermark which combine image features with SVD algorithm as well as the application algorithm in feature extraction based on improved SVD algorithm are put forward for the comparative analysis of corresponding principles and processing effects. The experimental results indicate that the content-based (instead of pixel-based) watermarking algorithm can better satisfy the perceptibility and robustness of digital watermarking, with huge application potentials and more development space