Search

In order to make full use of the joint sparse characteristics of sparse signals, a new distributed Subspace Pursuit algorithm based on the Subspace Pursuit (SP) algorithm is proposed in this paper. The algorithm improves the support set initialization and update processing of the traditional SP algorithm. During the initialization stage of the support set, the positions of the larger values of inner product appear most times were chosen as the initial set of support. In the update stage of the support set, sum coefficient values of all non-zero entries of signals and add the positions corresponding to the larger values to the support set. Compared with the existing SP algorithm and distributed algorithms which based on SP algorithm, the proposed method is capable of improving the accuracy of reconstruction, reducing the measured values and boosting the efficiency of the signal reconstruction. Theoretical analysis and simulation results show the effectiveness of the proposed method.

In order to suppress the range sidelobes and the grating lobes of the linear stepped frequency waveform, a non-linear stepped frequency (NSF) waveform design method by using range ambiguity function is proposed in this paper. Firstly, the reason of the sidelobes and the grating lobes occurring is analyzed based on range ambiguity function. Secondly, through compare the difference between the modified range ambiguity function, which has low sidelobes and grating lobes stepped frequency waveform, and the range ambiguity function which belongs to linear stepped frequency waveform, the designed frequencies of different sub-pluses can be obtained. In order to simplify the analysis process, due to the position and the amplitude of the sidelobes are periodic, then the non-linear frequencies design problem is transformed into a simple parameter estimation problem. At last, the frequencies can be quickly calculated through a simple formula. The simulation results show that the effectiveness of the proposed method.

An iteratively reweighted complex approximate message passing algorithm is proposed for cluster sparse inverse synthetic aperture radar imaging. The proposed algorithm improves the recovery performance of cluster sparse signal. A pattern-coupled weight update rule is introduced to establish a coupling relation between the sparsity patterns of neighboring coefficients. Real radar data results demonstrate the superior reconstruction performance in terms of reconstruction accuracy and computation time.

In this paper, a backstepping control design is presented for the longitudinal dynamics of a hypersonic flight vehicle. The GHV dynamics system is first decoupled into two subsystems, velocity subsystem and altitude subsystem. Then, the two subsystems are transformed into the strict-feedback forms, and on the basis of which the velocity controller and altitude controller are separately designed. Next, a comprehensive Lyapunov function based stability analysis is chosen to show that the tracking errors are semi-global uniformly bounded and the signals of the system are bounded. Numerical simulation results show that the proposed controller provides satisfied control performance.

On the basis of analysis of the principle of frequency control of motor speed, and on motor unified theory, the dynamic mathematical model of three phase asynchronous motor is established under three-phase coordinate transformation, To analyse this dynamic mathematical model, a simulation model is established under Matlab/Simulink and may simulation studies are performed to verily the feasibility of the proposed vector controlled drive. The results not only indicate the validity of the dynamic mathematical model and simulation model, but also show that by adjusting simulation parameters in each model, perfect simulation waveform is got.

In this paper, a new joint rapid method is presented, based on fast Fourier transformation (FFT) and multi-windows iterative smoothed pseudo Wigner-Ville distribution (SPWVD) to estimate parameters of frequency-hopping (FH) signals. Firstly, the approximate range including hop timing is searched by segmented FFT, then the both ends of the data segments by multi-windows iterative SPWVD, combining with statistics about the ends to obtain more accurate estimation of the parameters. This method can be easily implemented with lower computation complexity, and making real-time estimation possible. Simulation results show that the method can more rapidly and accurately estimate parameters of FH signals in the low SNR.

Under the extremely limited space experimental conditions, it's still a challenge for us to make the small sample of the data ensure the validity of experimental results in a certain confidence interval. We used the theory of small sample data analysis to design and synthesize strategies in order to be adequate for reasonably good knockdown of the number of samples required for the experiment. And we applied and verified the method to determine the number of operation sEMG signal samples by using an example of space ergonomics experiment. Finally, we analyzed the applicability and reliability of this method for series of experiments and looked forward to the trend of the research directions.

To effectively resolve the automatic recognition problem of analog measuring instruments, a reading recognition method based on improved Hough transformation is proposed. Firstly, the paper presents the image preprocessing operation. In order to make the image filtering better, an improved adaptive median filter method is adopted. Secondly, we explains in detail the recognition principle for the pointer angle using the improved Hough transform. And the pointer's reading is got using the linear relationship between scale and angle of instrument dial revolutionary. In order to further improve the recognition accuracy, as to the situation that pointer completely coincides with the whole tick, late correction will be done. Finally, the automatic recognition software is written in VC++6.0 IDE, and complete the recognition experiment for the AC voltmeter. The experimental results show that the method is very valid for the pointer angle recognition. In addition, the advantages of this study are as following: simple algorithm, good real-time and high precision.

Based on the theory of multi-body system and the motion of CNC gear hobbing machine, a volumetric error model for NC gear hobbing machine has been built, and the tooth surface equation is established, which includes original position errors and motion errors. Thus a preferable method is provided for carrying out real-time measurement and compensation, and further studying the effects that error factor takes on gear machining.

Regional comprehensive corridor takes on the characteristics of heavy transport demand and supply, and consists of many travel modes; meanwhile, competition is severe among different trip modes. In order to avoid vicious competition so that transportation market is disturbed, multi-objective and nonlinear programming model is built up, which is based on the maximization of unity of traveler. The objective of model is minimization of distance between practical capacity coefficient and ideal capacity coefficient. Multi-objective is transferred into single objective by TOPSIS method before model is solved. The transformed objective function is multi-peak function, so niche GA is introduced into solving the problem to obtain all of optimal solution. K-means clustering is adapted to classify the set of optimal solution. The centre of gravity of each cluster is taken as value of decision variables. At last a numeric example demonstrates the validity and feasibility.

The traditional CBIR (Content based image retrieval) global descriptor can't extract local features effectively, while the SIFT (Scale Invariant Feature Transform) algorithm only focus on local object features matching, and can not satisfy the need of fuzzy CBIR queries based on visual feeling. In view of this contradiction, an approach is proposed based on image salient region, in which the SIFT algorithm and the traditional CBIR color descriptors are combined, the entropy of salient point is used to distinguish the salient and non-salient regions of image. We use centroid weighted histogram and dominant color descriptor of MPEG-7 to describe the salient and non-salient region respectively, which not only effectively extract the local features of image objects, but also overcome the SIFT algorithm for high dimension, and vulnerable to interference with the global features. Experiments show that for general image retrieval application, the proposed method has better retrieval performance, with better robustness for geometric and perspective transform.

Image watermark resistant to geometric attacks is the focus and difficulty point in the current research on watermarking. This paper proposes a Wavelet-domain watermarking algorithm against geometric attacks based on the SUSAN feature points matching. The algorithm embeds watermarks into each sub-band of the wavelet coefficients of the host image with different strength adaptively adjusted in terms of the gray level of them. The choice of the embedding strategy effectively guarantees good transparency and strong robustness of the watermark. Furthermore, the paper improves the SUSAN operator, to extract the feature points robust to geometric distortion from the watermarked image. Before watermark detection, it uses the transformation of the corresponding points to correct geometric distortion, which ensures successful extraction of the watermark. Experiment results show that the proposed algorithm has good transparency, and it is also robust to common image processing and geometric attacks.

The development trend of gear hobbing machine which includes CNC, high-speed, high-accuracy, direct-drive and environment protection, is discussed firstly, and design principle of direct-drive gear hobbing machine of dry-cutting, which break through traditional structure design theory of gear cutting machine, is put forward. Direct-drive gear hobbing machine adopts the technologies of build-in motor spindle and DDR torque motor which directly drive the revolve of cutting tool and worktable, and completely cancels high precision gears and worm wheel in the machine transmission, so to realize that power source directly drive work parts of machine tool. The key technologies of direct-drive gear hobbing machine of dry-cutting, such as general layout technology, bed and chip removal system design technology, hob spindle and workpiece spindle structure design technology, are analyzed and studied in detail. It is of theoretic significance and practical value, for the study and development of direct drive gear-cutting machine of high-speed and high-precision.

The accurate motion compensation is important for gaining high resolution of imaging for some characteristics of missile-borne SAR, such as high speed, motion pose errors, high real-time and so on. The paper first introduces the two types of errors of missile-borne SAR and their effects to image. Subsequently, the first class of error, including normal face error and error along flight route, is described in details due to its serious effects to imaging. At the same time, the corresponding compensation algorithms are also showed. Finally, the paper gives the simulation results that the phase compensation to Doppler phase errors resulting from slant distance can improve the imaging resolution effectively. In the meantime, the phase compensation to Doppler modulation frequency errors resulting from speed error ameliorates Doppler mismatching. The resolution of the system reaches two meter.

To aim at the problem of blind detection of low-SNR direct-sequence spread spectrum (DS/SS) signals, this paper presents that the DS/SS signal has fractal characteristic, and uses the variance fractal dimension trajectory (VFDT) approach to detect the DS/SS signal contaminated by noise. The theoretic analysis and computer simulation results show that this algorithm can detect DS/SS signals at a lower SNR and runs more quickly than the usual detecting methods like the time-domain correlation detecting method, the cepstrum detecting method, and so on.

To aim at the problem of blind estimation of low-SNR DS/SS signals spreading sequence, an approach is proposed to estimate the spreading sequence from DS/SS signals in this paper. The method is based on neural network by a novel information criterion. The theoretic analysis and computer simulation results show that this algorithm has much less computation and storage than EVD, and its performance is better than the primitive neural network algorithm.

Base on existing maneuvering models and the IMM algorithm, this paper proposed one kind of IMM algorithm which used many kinds of maneuvering models for the high speed and highly maneuvering target tracking. Because of supplementary between different models, this algorithm has overcome some problems while using single model. The simulation results indicated, this algorithm has obtained a better balance between the performance and computational complexity, as well as very good practicability.