Search

Bearing failure is one of the foremost causes of breakdowns in rotating machinery and such failure can be catastrophic. Fault diagnosis is critical to maintaining the normal operation of the bearings. This paper proposes feature-level fusion method for rolling bearing fault diagnosis. Features are extracted from eight vibration signals to constitute a fusion vector. SVM is used for pattern recognition. The case study results show that the proposed method is useful for rolling bearing fault diagnosis.

For the kernel K-mean cluster method is run in an implicit feature space, the initial and iterative cluster centers cannot be defined explicitly. Against the deficiency of the initial cluster centers selected in the original space discretionarily in the existing methods, this paper proposes a new method for ensuring the clustering center that virtual clustering centers are defined in the feature space by the original classification as the initial cluster centers and the iteration clustering centers are ensured by the further virtual classification. The improved method is used for fault diagnosis of roller bearing that achieves a good cluster and diagnosis result, which demonstrates the effectiveness of the proposed method.

Multisensor information fusion, when applied to fault diagnosis, the time-space scope, and the quantity of information are expanded compared to what could be acquired by a single sensor, so the diagnostic object can be described more comprehensively. This paper presents a methodology of fault diagnosis in rotating machinery using multisensor information fusion that all the features are calculated using vibration data in time domain to constitute fusional vector and the support vector machine (SVM) is used for classification. The effectiveness of the presented methodology is tested by three case studies: diagnostic of faulty gear, rolling bearing, and identification of rotor crack. For each case study, the sensibilities of the features are analyzed. The results indicate that the peak factor is the most sensitive feature in the twelve time-domain features for identifying gear defect, and the mean, amplitude square, root mean square, root amplitude, and standard deviation are all sensitive for identifying gear, rolling bearing, and rotor crack defect comparatively.

With the development of signal processing in sonar technology, the dynamic range of signal data is expanded. The traditional visual method of gray scale cannot meet the demand of data display. Pseudo color Transformation combined with human vision response is an effective method to improve the capability of sonar operators in discriminating targets and to improve their reaction speed. A pseudo-color coding based on CIELab space is designed for sonar visual system. Sea trial data demonstrate that the method provides more target detail information, which is of high value in sonar application.

Tests of rodent visual capacities are typically performed under standard laboratory illumination. However, light level can have subtle and complex effects on behavior in rodents. We tested Mongolian gerbils (Meriones unguiculatus) - a species that shows individual differences in activity pattern - on visual tasks at three ecologically relevant levels of ambient illuminance: approximating moonlight (1 lx), dawn or dusk (10 lx), and low daylight (100 lx). A jumping task and a grating discrimination quantified depth perception and grating acuity, respectively. Illuminance variations had important effects on behavior. Gerbils jumped faster in lower light on the depth discrimination task, but were also less accurate than in bright light. Daytime activity levels (possibly reflecting variations in activity pattern) mediated these effects, with animals that were awake during a lower proportion of daytime hours exhibiting a trend toward jumping faster with lower light level as compared to more day-active gerbils. Moreover, while illuminance did not affect acuity overall, it interacted with activity levels in a complex way: in both the darkest and lowest light levels, animals that were awake during a greater proportion of daytime hours showed higher acuity levels than animals that were less active during the day. These results indicate that gerbils show behavioral profiles in vision tasks that represent an interaction between visual ability and illuminance-sensitive motivational or emotional actors, perhaps including chronotype. Thus, the most ecologically relevant assessments of the visual performance of rodents will likely be achieved by testing at species-specific preferred light levels.

Automatic modulation recognition of modulation signals is the key problem in non-cooperative communication systems. The method which combines the rough set theory and the neural network is designed for identifying the six modulation types based on the research on the feature set of digital modulation recognition. Simulation results show that the new method simplifies the structure of neural networks and decreases the training time without reducing the recognition rate.

On automatic modulation there are two approaches, decision-theoretic and statistical pattern. An automatic modulation recognition system to recognize four digital signal classes as: MASK, MFSK, MPSK, MQAM is proposed in this paper, which using decision-theoretic based feature set addition to statistical pattern based feature set with VLBP(variable learning rate back-propagation)BP neural network and Bayesian normalized BP neural network. Performance is generally good when Signal to Noise Ratios (SNR) in 0-10dB, simulations show the results even larger than 95%, that confirm the robustness and practicality of this recognition method.

This article studies the classical MDS and dwMDS location algorithm. On this basis, the relaxation iteration method is introduced to replace SMACOF algorithm as optimization objective function. The results showed that the network connectivity increases with the increased communication radius, and it has a good effect that the relaxation method of positioning results in the node density from 2 to 4. The positioning performance of the relaxation method is even better than dwMDS algorithm. That is the algorithm for sparse node distribution network. And when the node density is too large the algorithm's maximum number of iterations should be limited.

Along with the profound development of wireless sensor network, and in the environment of complicated and full of counteract communication, correct localization information of node is the prerequisite and basis of accomplishing some modules, such as construction of sensor network, maintenance, incident localization, target tracking. Thus, localization security becomes a key technique. Mastering the merits of wireless sensor network attack is a vital prerequisite of solving security question. The paper mainly discusses attack module of localization system in WSN, and makes a specific research on attack theory, attack features, attack effect, etc. On the basis of the attack module analysis, the paper comes up with localization security demand.

Change the BP Algorithm rely on the gradient information to adjust the network weights, the optimal gene sequences decode the expression tree so get the best neural network structure and the evolution of weights and thresholds, so that the structure of artificial neural network and weights can be optimized at the same time. This method gives the neural network mapping capabilities and GEP's ability to solve complex problems, accelerate the learning speed of the network, improve the approximation ability and generalization ability. GEP-BP will be used in Shanghai Composite Index forecast. Experimental results show that this method improved the prediction accuracy and achieved a better prediction.

The paper presents a simple cognitive UWB system model with queuing mechanism. In accordance with this model a good spectrum was obtained by tuning transmitter parameters adaptively. It can support dynamic avoidance of narrow-band interference. While implementing a true cognitive UWB radio is still quite challenging today because of the limitations of current digital processing technologies, intelligent spectrum-sensing capabilities and dynamic waveform generation algorithms. So our next-step work is to find better pulse design methods and algorithms for acquiring a perfect cognitive UWB system.

Blind source separation problem have recently drawn a lot of attention in unsupervised neural learning. In the current approaches, the additive noise is negligible so that it can be omitted from the consideration. To be applicable in realistic scenarios, blind source separation approaches should deal evenly with the presence of noise. In this paper, a method is proposed of combining wavelet threshold de-noising and independent component analysis to the blind source separation problem for mixing images corrupter with white noise. We first use wavelet threshold to de-noise and then use a new blind separation algorithm of FASTICA to separate the wavelet de-noised images. The result shows that this method may reduce the affect of noise and improve the signal-noise ratio (SNR) of separation images, accordingly renew the original images.

This paper introduces the principle of the time-recording on automobile driver training, discusses an error tolerance plan of data storage countering some possible inferences of the system and analyses of the double-layered (16, 11, 4) minimal optimal odd-weight error-correcting code. From the analysis, one can conclude that this code has the merit of low redundancy and strong ability of anti-inference and detecting and correcting error.

The DSP-based parameters measuring instrumentation of low-voltage power net utilizes the sampling value of voltage and current to calculate the parameters of the power net. This paper puts forward a new sampling method that is based on the amplitude error. It fully utilizes the physical meaning of Fourier series' coefficients and the high-speed calculation merits of DSP. The realization of the algorithm is easier on DSP than on other single-chip microcomputers.

This paper presents the design of a helix filter used to suppress the jamming in a microwave remote-vision system, analyses the mathematical model of the helix filter and gives the design of the helix filter with a frequency bandwidth from 566 MHz to 578 MHz.