ISSN 
2010 Vol. 39, No. 6
2010, 39(6): 803-809.
doi: 10.3969/j.issn.1001-0548.2010.06.001
Abstract:
The research progress and applications of functional networks (FNt) models in recent years are reviewed. At first, according to the functional network model structure characteristics, FNt models are classified into two typical types in accordance with the functional networks. Secondly, recent applications of FNt are introduced in many fields, such as times series analysis, differential equations, CAD, nonlinear regression, numerical optimize computation, nonlinear system identification, detection and prediction, complex system modeling, etc. Finally, some remarks on the further research and directions of FNt are presented.
The research progress and applications of functional networks (FNt) models in recent years are reviewed. At first, according to the functional network model structure characteristics, FNt models are classified into two typical types in accordance with the functional networks. Secondly, recent applications of FNt are introduced in many fields, such as times series analysis, differential equations, CAD, nonlinear regression, numerical optimize computation, nonlinear system identification, detection and prediction, complex system modeling, etc. Finally, some remarks on the further research and directions of FNt are presented.
2010, 39(6): 810-816.
doi: 10.3969/j.issn.1001-0548.2010.06.002
Abstract:
The problem of impulsive stabilization of delayed cellular neural networks (DCNNs) via partial states is discussed. The time delay is allowed to be time-varying. By utilizing the piecewise linear property of the activation function of DCNNs and applying piecewise differential Lyapunov combined with Razumikhin- type analysis techniques, a sufficient condition for the existence of the impulsive control law via partial states is derived. The sufficient condition is given in terms of linear matrix inequalities concerning the interconnection matrices and the bounds of the impulsive intervals. By using this result, an impulsive stabilization scheme for a class of DCNNs is proposed. The impulsive stabilization scheme only utilizes the output of partial states of the controlled DCNN. A numerical example illustrates the efficiency of the proposed method.
The problem of impulsive stabilization of delayed cellular neural networks (DCNNs) via partial states is discussed. The time delay is allowed to be time-varying. By utilizing the piecewise linear property of the activation function of DCNNs and applying piecewise differential Lyapunov combined with Razumikhin- type analysis techniques, a sufficient condition for the existence of the impulsive control law via partial states is derived. The sufficient condition is given in terms of linear matrix inequalities concerning the interconnection matrices and the bounds of the impulsive intervals. By using this result, an impulsive stabilization scheme for a class of DCNNs is proposed. The impulsive stabilization scheme only utilizes the output of partial states of the controlled DCNN. A numerical example illustrates the efficiency of the proposed method.
2010, 39(6): 817-819,830.
doi: 10.3969/j.issn.1001-0548.2010.06.003
Abstract:
In array signal processing, the cases of mixed circular and non-circular sources are more and more frequently encountered in complex signal situation. By analyzing the subspace property of the mixed circular and non-circular signal sources in detail, an enhanced and fast method for direction of arrival (DOA) estimation and identification of mixed circular and non-circular sources is proposed. Since the proposed method efficiently utilizes the useful non-circularity properties of the mixed circular and non-circular source model, it has better performance of DOA estimation compared with the method only based on homogeneous source model and its performance is closed to the theoretical bound. Simulation results show that by using the proposed method in DOA estimation, the number of mixed signal sources can be larger than the number of array elements.
In array signal processing, the cases of mixed circular and non-circular sources are more and more frequently encountered in complex signal situation. By analyzing the subspace property of the mixed circular and non-circular signal sources in detail, an enhanced and fast method for direction of arrival (DOA) estimation and identification of mixed circular and non-circular sources is proposed. Since the proposed method efficiently utilizes the useful non-circularity properties of the mixed circular and non-circular source model, it has better performance of DOA estimation compared with the method only based on homogeneous source model and its performance is closed to the theoretical bound. Simulation results show that by using the proposed method in DOA estimation, the number of mixed signal sources can be larger than the number of array elements.
2010, 39(6): 820-825.
doi: 10.3969/j.issn.1001-0548.2010.06.004
Abstract:
A blind method for symbol rate estimation of MPSK/MQAM signals by using the nonlinear process is presented. Utilizing the phase changing between inter-symbols and through the nonlinear transforming of zero IF signals, the spectral lines corresponding to the symbol rate can be extracted. The fourth-order cumulant is used for suppressing the strong background noise of symbol rate spectral lines in this paper. This method is robust to Gaussian noise and can achieve good estimation results with lower SNR. At the same time, it has lower computational complexity and is applicable to different shaping filters and different modulation orders of MPSK/MQAM signals.
A blind method for symbol rate estimation of MPSK/MQAM signals by using the nonlinear process is presented. Utilizing the phase changing between inter-symbols and through the nonlinear transforming of zero IF signals, the spectral lines corresponding to the symbol rate can be extracted. The fourth-order cumulant is used for suppressing the strong background noise of symbol rate spectral lines in this paper. This method is robust to Gaussian noise and can achieve good estimation results with lower SNR. At the same time, it has lower computational complexity and is applicable to different shaping filters and different modulation orders of MPSK/MQAM signals.
2010, 39(6): 826-830.
doi: 10.3969/j.issn.1001-0548.2010.06.005
Abstract:
The reasons for the appearance of subliminal channels and their applications in digital signature schemes are analyzed. The wideband and narrowband subliminal channels in the Schnorr signature are discussed. And a new subliminal-free protocol is designed. In the new protocol, the warden participates in the generation of session keys in order to guarantee their randomness and privacy. It is shown that the protocol can completely close the subliminal channels existing in the random session keys in the Schnorr signature scheme. In addition, the signature authority of the signer is guaranteed. The security of the proposed protocol is based on both the discrete logarithm intractability assumption and the existence of collision-free hash functions. To generate a signature, it only needs to perform one modular exponentiation for each of the signer and the warden.
The reasons for the appearance of subliminal channels and their applications in digital signature schemes are analyzed. The wideband and narrowband subliminal channels in the Schnorr signature are discussed. And a new subliminal-free protocol is designed. In the new protocol, the warden participates in the generation of session keys in order to guarantee their randomness and privacy. It is shown that the protocol can completely close the subliminal channels existing in the random session keys in the Schnorr signature scheme. In addition, the signature authority of the signer is guaranteed. The security of the proposed protocol is based on both the discrete logarithm intractability assumption and the existence of collision-free hash functions. To generate a signature, it only needs to perform one modular exponentiation for each of the signer and the warden.
2010, 39(6): 831-834.
doi: 10.3969/j.issn.1001-0548.2010.06.006
Abstract:
Based on algebraic attack, a new reconstruction method of Boolean functions from the partial truth table is proposed. For the Boolean function with n variables and the degree of d, the proposed method requires O(N) values in the truth table, and the computational complexity is O(N3), and the memory complexity is O(N), where N=1+C1n+C2n+…+Cdn. From the above complexity, the lower the degree of the Boolean function is，the more efficient the method is. The proposed attack shows the designer of stream cipher should use Boolean functions with low degree carefully.
Based on algebraic attack, a new reconstruction method of Boolean functions from the partial truth table is proposed. For the Boolean function with n variables and the degree of d, the proposed method requires O(N) values in the truth table, and the computational complexity is O(N3), and the memory complexity is O(N), where N=1+C1n+C2n+…+Cdn. From the above complexity, the lower the degree of the Boolean function is，the more efficient the method is. The proposed attack shows the designer of stream cipher should use Boolean functions with low degree carefully.
2010, 39(6): 835-840.
doi: 10.3969/j.issn.1001-0548.2010.06.007
Abstract:
In this paper, a blind PTS scheme in MC-CDMA systems is proposed. Based on the relationship between the dispreading codes and side information, the new scheme utilizes the phase recovering with the partial dispread signal, and makes the optimal decision based on minimum power error of the dispread signals. Simulation shows that compared with the conventional PTS scheme, the proposed scheme can recover the phase information without SI at the receiver and provide better frequency efficiency and transmitting performance.
In this paper, a blind PTS scheme in MC-CDMA systems is proposed. Based on the relationship between the dispreading codes and side information, the new scheme utilizes the phase recovering with the partial dispread signal, and makes the optimal decision based on minimum power error of the dispread signals. Simulation shows that compared with the conventional PTS scheme, the proposed scheme can recover the phase information without SI at the receiver and provide better frequency efficiency and transmitting performance.
2010, 39(6): 841-844,849.
doi: 10.3969/j.issn.1001-0548.2010.06.008
Abstract:
In this paper, bandpass sampling and subspace reconstruction methods of ultra-wideband (UWB) signals are proposed, which can greatly reduce the sampling rate in UWB communications. The UWB signals are passed through a bandpass filter and sampled at several times of the innovation rate. Then the amplitudes and time shifts of UWB signals can be obtained by subspace reconstruction method. Analysis and simulation results show that the subspace method can precisely reconstruct the UWB signals at a sampling rate which is much lower than Nyquist rate.
In this paper, bandpass sampling and subspace reconstruction methods of ultra-wideband (UWB) signals are proposed, which can greatly reduce the sampling rate in UWB communications. The UWB signals are passed through a bandpass filter and sampled at several times of the innovation rate. Then the amplitudes and time shifts of UWB signals can be obtained by subspace reconstruction method. Analysis and simulation results show that the subspace method can precisely reconstruct the UWB signals at a sampling rate which is much lower than Nyquist rate.
2010, 39(6): 845-849.
doi: 10.3969/j.issn.1001-0548.2010.06.009
Abstract:
The packet loss differentiating mechanism of TCP for heterogeneous networks plays an important role in the stability of wireless networks, relative one-way delay (ROD) of packet pair is used as the differentiating parameter to distinguish the loss type. The accuracy of this algorithm depends on ROD samples membership functions and parameters estimation. In order to differentiate the packet loss pattern more accurately, the initialization method based on potential functions is proposed by analyzing the traditional expectation maximization (EM) algorithm in Gaussian mixture model. Then the improved EM (PEM) algorithm is applied in the construction of different packet loss membership functions in the situation when the network congestion and wireless error coexist. The simulation results indicate that this algorithm has better convergence characteristics and stability, and has well building effect in the construction of different packet loss pattern membership functions.
The packet loss differentiating mechanism of TCP for heterogeneous networks plays an important role in the stability of wireless networks, relative one-way delay (ROD) of packet pair is used as the differentiating parameter to distinguish the loss type. The accuracy of this algorithm depends on ROD samples membership functions and parameters estimation. In order to differentiate the packet loss pattern more accurately, the initialization method based on potential functions is proposed by analyzing the traditional expectation maximization (EM) algorithm in Gaussian mixture model. Then the improved EM (PEM) algorithm is applied in the construction of different packet loss membership functions in the situation when the network congestion and wireless error coexist. The simulation results indicate that this algorithm has better convergence characteristics and stability, and has well building effect in the construction of different packet loss pattern membership functions.
2010, 39(6): 850-853,858.
doi: 10.3969/j.issn.1001-0548.2010.06.010
Abstract:
Nonuniform sample of SAR signal will add the noise into the signals, mix the spectrum of the echo signals with the spectrum of the noises, increase ambiguity of the signals, and depress the SAR imaging quality. In wide swath SAR systems, the multi-channel receivers are used in Azimuth for high resolution image, the nonuniform sampling will be meted. This paper carefully analyzes the spectrum ambiguity of the nonuniform sampling SAR signal and reconstructs unambiguous spectrum of the signal by periodic time offset spectrum reconstruction way; Based on the difference and correlation between the multi-channel signals and single channel signal, we have finished the multi-channel SAR imaging by CS algorithm, the simulation results prove the correctness of theoretic analysis.
Nonuniform sample of SAR signal will add the noise into the signals, mix the spectrum of the echo signals with the spectrum of the noises, increase ambiguity of the signals, and depress the SAR imaging quality. In wide swath SAR systems, the multi-channel receivers are used in Azimuth for high resolution image, the nonuniform sampling will be meted. This paper carefully analyzes the spectrum ambiguity of the nonuniform sampling SAR signal and reconstructs unambiguous spectrum of the signal by periodic time offset spectrum reconstruction way; Based on the difference and correlation between the multi-channel signals and single channel signal, we have finished the multi-channel SAR imaging by CS algorithm, the simulation results prove the correctness of theoretic analysis.
2010, 39(6): 854-858.
doi: 10.3969/j.issn.1001-0548.2010.06.011
Abstract:
Adaptive pulse compression technique is an efficient technology to mitigate the effects of side-lobes. In this paper, a modified adaptive pulse compression algorithm is proposed, which utilizes diagonal loading technology to compress noise, improve signal-to-noise ratio, and enhance the capability of suppressing the side-lobes. The simulation results manifest that the modified adaptive pulse compression is superior to the conventional adaptive pulse compression algorithm.
Adaptive pulse compression technique is an efficient technology to mitigate the effects of side-lobes. In this paper, a modified adaptive pulse compression algorithm is proposed, which utilizes diagonal loading technology to compress noise, improve signal-to-noise ratio, and enhance the capability of suppressing the side-lobes. The simulation results manifest that the modified adaptive pulse compression is superior to the conventional adaptive pulse compression algorithm.
2010, 39(6): 859-863.
doi: 10.3969/j.issn.1001-0548.2010.06.012
Abstract:
Chaotic sequences have been widely used as pseudorandom sequences. But the problem of how to judge the performances of their autocorrelation functions, up to now, has not yet been solved completely. Because of this, the applications of chaotic sequences are limited. In the paper, by method of phase space, we prove that the autocorrelation performance of a chaotic sequence is determined by whether its phase space trajectory is axis symmetrical, and we deduce a theorem that a sequence which phase space trajectory is axis symmetrical has good autocorrelation performance. The theorem is verified by simulations.
Chaotic sequences have been widely used as pseudorandom sequences. But the problem of how to judge the performances of their autocorrelation functions, up to now, has not yet been solved completely. Because of this, the applications of chaotic sequences are limited. In the paper, by method of phase space, we prove that the autocorrelation performance of a chaotic sequence is determined by whether its phase space trajectory is axis symmetrical, and we deduce a theorem that a sequence which phase space trajectory is axis symmetrical has good autocorrelation performance. The theorem is verified by simulations.
2010, 39(6): 864-869.
doi: 10.3969/j.issn.1001-0548.2010.06.013
Abstract:
In tradition, the diagnosis of delamination in either grotto wall paintings or palace wall paintings is achieved by identifying the sound when lightly knocking on the wall by hand, such experience is practical in determining the area and degree of delamination, but it depends on subjective sensation, lacking some theoretical principles. Focusing on the propagation of high frequency pulse electromagnetic waves in layered lossy and dispersive medium and after the physical forward modeling experiment, this paper tries to locate delamination in polished wall paintings in Tibet by wall coupling antennas using RAMAC ground penetrating radar. It is shown that the ultra-wide band ground penetrating radar is capable of detecting delamination in vertical resolution of about 5 mm when it is equipped with a transmitting antenna of 1.6 GHz central frequency. It is a quantitative attempt to detect delamination in wall paintings, the research results promote technical improvement for conservation of wall paintings.
In tradition, the diagnosis of delamination in either grotto wall paintings or palace wall paintings is achieved by identifying the sound when lightly knocking on the wall by hand, such experience is practical in determining the area and degree of delamination, but it depends on subjective sensation, lacking some theoretical principles. Focusing on the propagation of high frequency pulse electromagnetic waves in layered lossy and dispersive medium and after the physical forward modeling experiment, this paper tries to locate delamination in polished wall paintings in Tibet by wall coupling antennas using RAMAC ground penetrating radar. It is shown that the ultra-wide band ground penetrating radar is capable of detecting delamination in vertical resolution of about 5 mm when it is equipped with a transmitting antenna of 1.6 GHz central frequency. It is a quantitative attempt to detect delamination in wall paintings, the research results promote technical improvement for conservation of wall paintings.
2010, 39(6): 870-874,890.
doi: 10.3969/j.issn.1001-0548.2010.06.014
Abstract:
Landing gear of aircraft is the key component to ensure safety during the process of takeoff and landing. This paper analyzes the dynamic characteristics of the landing gear system. Because of the inclusion of MR fluid, the landing gear damper system becomes a nonlinear system. In order to study the characteristics of the system, this paper builds the mathematical model of the landing gear damper by linearizing the nonlinear system. Based on the model, the frequency response of aircraft body displacement, aircraft body acceleration, landing gear displacement, and tire dynamic loading of the landing gear are analyzed. And the time-domain response of random road excitations is also analyzed. These analyses would be helpful in structure designing of the landing gear with MR damper and the advanced control of the damper system.
Landing gear of aircraft is the key component to ensure safety during the process of takeoff and landing. This paper analyzes the dynamic characteristics of the landing gear system. Because of the inclusion of MR fluid, the landing gear damper system becomes a nonlinear system. In order to study the characteristics of the system, this paper builds the mathematical model of the landing gear damper by linearizing the nonlinear system. Based on the model, the frequency response of aircraft body displacement, aircraft body acceleration, landing gear displacement, and tire dynamic loading of the landing gear are analyzed. And the time-domain response of random road excitations is also analyzed. These analyses would be helpful in structure designing of the landing gear with MR damper and the advanced control of the damper system.
2010, 39(6): 875-879,890.
doi: 10.3969/j.issn.1001-0548.2010.06.015
Abstract:
According to the needs of conditional-based maintenance for electronic system, the paper presents a novel method to monitor transmitter's state and evaluate its health. Through analyzing transmitter's key signal, the corresponding wavelet features can be extracted and are further sent to CHMM used as the state monitor to calculate the Kullback-Leibler (KL) distance of unknown state, and thus, transmitter's incipient fault process can be converted into the obvious change of KL distance successfully. Based on the KL distance, transmitter's health can be correctly evaluated and it becomes easy to decide to repair it or not. Finally the tested performance of the proposed method indicates that the method is effective and provides basis for condition-based maintenance.
According to the needs of conditional-based maintenance for electronic system, the paper presents a novel method to monitor transmitter's state and evaluate its health. Through analyzing transmitter's key signal, the corresponding wavelet features can be extracted and are further sent to CHMM used as the state monitor to calculate the Kullback-Leibler (KL) distance of unknown state, and thus, transmitter's incipient fault process can be converted into the obvious change of KL distance successfully. Based on the KL distance, transmitter's health can be correctly evaluated and it becomes easy to decide to repair it or not. Finally the tested performance of the proposed method indicates that the method is effective and provides basis for condition-based maintenance.
2010, 39(6): 880-885.
doi: 10.3969/j.issn.1001-0548.2010.06.016
Abstract:
An actual measurement model of star sensor is established by geometrical method. The effect of relative factors to the actual imaging location of navigation star is considered into the ideal measurement model of star sensor. These factors include focal length deviation, optical lens distortion, image sensor tilt, image sensor rotate, primary point bias, etc. The affecting law of the factors on angle-measuring accuracy of star sensor is analyzed. The star sensor is calibrated through the measured calibration data and the model parameters solved by least square method. The experimental results show that the angle measurement accuracy of x and y directions is 1.6″ and 2.3″ after calibration, but 19.6″ and 17.8″ before calibration.
An actual measurement model of star sensor is established by geometrical method. The effect of relative factors to the actual imaging location of navigation star is considered into the ideal measurement model of star sensor. These factors include focal length deviation, optical lens distortion, image sensor tilt, image sensor rotate, primary point bias, etc. The affecting law of the factors on angle-measuring accuracy of star sensor is analyzed. The star sensor is calibrated through the measured calibration data and the model parameters solved by least square method. The experimental results show that the angle measurement accuracy of x and y directions is 1.6″ and 2.3″ after calibration, but 19.6″ and 17.8″ before calibration.
2010, 39(6): 886-890.
doi: 10.3969/j.issn.1001-0548.2010.06.017
Abstract:
A novel corner detection algorithm based on regulated morphology is proposed in this paper. Symmetric disk structure element is adopted in the algorithm. Firstly, by choosing a proper regulated parameter, the image is processed with a series of operations such as regulated erosion and regulated dilation to get accurate positions of the convex and concave corners respectively, and then all corners in the image are obtained by combing the two kinds of corners together. The use of disk structure element highly decreases the computation cost. The method based on regulated morphology is more precise than methods based on traditional morphology. The experiment results prove that when a proper value of the regulated parameter is chosen, the proposed new detector can precisely detect corners in images.
A novel corner detection algorithm based on regulated morphology is proposed in this paper. Symmetric disk structure element is adopted in the algorithm. Firstly, by choosing a proper regulated parameter, the image is processed with a series of operations such as regulated erosion and regulated dilation to get accurate positions of the convex and concave corners respectively, and then all corners in the image are obtained by combing the two kinds of corners together. The use of disk structure element highly decreases the computation cost. The method based on regulated morphology is more precise than methods based on traditional morphology. The experiment results prove that when a proper value of the regulated parameter is chosen, the proposed new detector can precisely detect corners in images.
2010, 39(6): 891-895.
doi: 10.3969/j.issn.1001-0548.2010.06.018
Abstract:
A method to simplify the model for forecasting the soil moisture of tobacco field by aiming at the same growing stage and area is proposed and a model to forecast the soil moisture of tobacco field based on BP neural network is established, in which the temperature, sunlight, rainfall, evaporation, and initial moisture of soil at a stage beginning are taken as the input and the soil moisture of tobacco field as the output. The research result shows that the prediction model of soil moisture has a good accuracy and has a wide range of adaptability.
A method to simplify the model for forecasting the soil moisture of tobacco field by aiming at the same growing stage and area is proposed and a model to forecast the soil moisture of tobacco field based on BP neural network is established, in which the temperature, sunlight, rainfall, evaporation, and initial moisture of soil at a stage beginning are taken as the input and the soil moisture of tobacco field as the output. The research result shows that the prediction model of soil moisture has a good accuracy and has a wide range of adaptability.
2010, 39(6): 896-899,910.
doi: 10.3969/j.issn.1001-0548.2010.06.019
Abstract:
Aiming at the problems of low efficiency in service discovery of traditional web service, a similarity-based semantic web service discovery model is proposed, with the application of semantic web technology and ontology. The mechanism uses Ontology Web Language for Service (OWL_S) to describe web service and publishes service with semantic information. Ontology reasoning is added into service searching to measure its degree of functional matching and reputation value. The experiments suggest that the proposed method is better than that of Universal Description Discovery and Integration (UDDI) in precision and recall ratio.
Aiming at the problems of low efficiency in service discovery of traditional web service, a similarity-based semantic web service discovery model is proposed, with the application of semantic web technology and ontology. The mechanism uses Ontology Web Language for Service (OWL_S) to describe web service and publishes service with semantic information. Ontology reasoning is added into service searching to measure its degree of functional matching and reputation value. The experiments suggest that the proposed method is better than that of Universal Description Discovery and Integration (UDDI) in precision and recall ratio.
2010, 39(6): 900-905.
Abstract:
A new algorithm of wavelet sampling is proposed since the typical wavelet sampling lacks the capacity to filter the continuous signal by the samples as Shannon sampling does. From the viewpoint of Multiresolution approximation, a new cost function based on the samples is introduced to estimate the approximation of signal in this algorithm, so that the signal is decomposed into the orthogonal components, which has break through the limit of typical wavelet sampling that only considers the construction of signal in one Hilbert space.
A new algorithm of wavelet sampling is proposed since the typical wavelet sampling lacks the capacity to filter the continuous signal by the samples as Shannon sampling does. From the viewpoint of Multiresolution approximation, a new cost function based on the samples is introduced to estimate the approximation of signal in this algorithm, so that the signal is decomposed into the orthogonal components, which has break through the limit of typical wavelet sampling that only considers the construction of signal in one Hilbert space.
2010, 39(6): 906-910.
doi: 10.3969/j.issn.1001-0548.2010.06.021
Abstract:
A certificateless public key encryption (CL-PKE) algorithm is presented. The proposed CL-PKE algorithm is based on the nature BDHP difficulty assumption and therefore avoids paring computation on elliptic curves, which is the most expensive operation in the encryption algorithm. In CL-PKE algorithm, the selective-ID model is applied instead of the random oracle model. The security and efficiency of the algorithm can be improved compared with some other CL-PKE schemes.
A certificateless public key encryption (CL-PKE) algorithm is presented. The proposed CL-PKE algorithm is based on the nature BDHP difficulty assumption and therefore avoids paring computation on elliptic curves, which is the most expensive operation in the encryption algorithm. In CL-PKE algorithm, the selective-ID model is applied instead of the random oracle model. The security and efficiency of the algorithm can be improved compared with some other CL-PKE schemes.
2010, 39(6): 911-914,919.
doi: 10.3969/j.issn.1001-0548.2010.06.022
Abstract:
To improve the efficiency of the forensic analysis method on data mining, this paper proposes a new method for the forensic analysis of the behavior profiling on the longest frequent pattern which is constructed by immune clonal algorithm. Taking the behavior data and the candidate pattern of the frequent item sets as the antigen and the antibody respectively, the support of the antigen to the antibody as the function of affinity, the key attribute as the constraint condition, and the minimal support as the screening condition, the behavior profiling on the longest frequent pattern is built with the help of the immune clonal operation to antibody. The abnormal data are detected by the matching method that the audit data pass through the list items of the behavior profiling. The proposed method and the method on Apriori-CGA are applied in the same problem. The comparison results indicate that the setting up time of behavior profiling and the test time of abnormal data are dramaticly reduced. Therefore, the proposed method has a good ability in the efficiency of forensic analysis and electronic crime investigation.
To improve the efficiency of the forensic analysis method on data mining, this paper proposes a new method for the forensic analysis of the behavior profiling on the longest frequent pattern which is constructed by immune clonal algorithm. Taking the behavior data and the candidate pattern of the frequent item sets as the antigen and the antibody respectively, the support of the antigen to the antibody as the function of affinity, the key attribute as the constraint condition, and the minimal support as the screening condition, the behavior profiling on the longest frequent pattern is built with the help of the immune clonal operation to antibody. The abnormal data are detected by the matching method that the audit data pass through the list items of the behavior profiling. The proposed method and the method on Apriori-CGA are applied in the same problem. The comparison results indicate that the setting up time of behavior profiling and the test time of abnormal data are dramaticly reduced. Therefore, the proposed method has a good ability in the efficiency of forensic analysis and electronic crime investigation.
2010, 39(6): 915-919.
doi: 10.3969/j.issn.1001-0548.2010.06.023
Abstract:
This paper proposes a new approach to perform the discriminant analysis on the labelled high dimensional image data with intra-class sub-manifolds. Real world images are usually taken from the different camera views. Pose, illumination, glasses and gender of the persons taking the facial images usually lead to multi-modality or high curvature of the underlying manifold structures. These variations result in the degraded performance of many existing algorithms. This paper proposes to preserve the within-class local structure, while imposing constrain on the variances only in the directions normal to the between-class margin. The experiments on Yale-B and UMIST face database show that the proposed algorithm outperforms many approaches such as LPP (locality preserving projections) and FDA (fisher discriminant analysis).
This paper proposes a new approach to perform the discriminant analysis on the labelled high dimensional image data with intra-class sub-manifolds. Real world images are usually taken from the different camera views. Pose, illumination, glasses and gender of the persons taking the facial images usually lead to multi-modality or high curvature of the underlying manifold structures. These variations result in the degraded performance of many existing algorithms. This paper proposes to preserve the within-class local structure, while imposing constrain on the variances only in the directions normal to the between-class margin. The experiments on Yale-B and UMIST face database show that the proposed algorithm outperforms many approaches such as LPP (locality preserving projections) and FDA (fisher discriminant analysis).
2010, 39(6): 920-923.
Abstract:
Iris localization is the most critical step of an iris identification system. Since it determines the accuracy and efficiency of the system. To improve its accuracy and efficiency, we proposed a new iris localization algorithm based on Multi-resolution analysis to detect the edge points of iris at one appropriate resolution and set ellipses to these edge points. The most important merit of this algorithm is that it is seldom affected by lash occlusion problem and can localize the off-angle iris image.
Iris localization is the most critical step of an iris identification system. Since it determines the accuracy and efficiency of the system. To improve its accuracy and efficiency, we proposed a new iris localization algorithm based on Multi-resolution analysis to detect the edge points of iris at one appropriate resolution and set ellipses to these edge points. The most important merit of this algorithm is that it is seldom affected by lash occlusion problem and can localize the off-angle iris image.
2010, 39(6): 924-928,951.
doi: 10.3969/j.issn.1001-0548.2010.06.025
Abstract:
Based on the 6LowPAN architecture, a design scheme on next-generation all-IP wireless sensor network is proposed. The scheme creates an IPv6 address structure, an automatic IPv6-address configuration algorithm and routing algorithm. In addition, the scheme also includes a fragmentation algorithm on the adaption layer and a reduced IPv6 protocol stack in wireless sensor networks. This scheme is implemented in both a physical experimental enviroment and a simulation environment respectively, and its performance is analyzed. The experimental data prove the validity and correctness of the scheme.
Based on the 6LowPAN architecture, a design scheme on next-generation all-IP wireless sensor network is proposed. The scheme creates an IPv6 address structure, an automatic IPv6-address configuration algorithm and routing algorithm. In addition, the scheme also includes a fragmentation algorithm on the adaption layer and a reduced IPv6 protocol stack in wireless sensor networks. This scheme is implemented in both a physical experimental enviroment and a simulation environment respectively, and its performance is analyzed. The experimental data prove the validity and correctness of the scheme.
2010, 39(6): 929-933.
doi: 10.3969/j.issn.1001-0548.2010.06.026
Abstract:
For the intension of buffering resources in network-on-chip (NoC), a buffer allocation algorithm is proposed. Given buffering space budget, our algorithm automatically allocates the resources on each input channel, in different routers across the chip, to match the traffic load, such that the overall performance is maximized. In the algorithm, a novel analytical model for adaptive routing is used to quickly detect potential performance bottlenecks in the system. Simulation results indicate that our algorithm can get lower average packet latency than uniform allocation strategy, and about 33% savings in buffering resources can be achieved.
For the intension of buffering resources in network-on-chip (NoC), a buffer allocation algorithm is proposed. Given buffering space budget, our algorithm automatically allocates the resources on each input channel, in different routers across the chip, to match the traffic load, such that the overall performance is maximized. In the algorithm, a novel analytical model for adaptive routing is used to quickly detect potential performance bottlenecks in the system. Simulation results indicate that our algorithm can get lower average packet latency than uniform allocation strategy, and about 33% savings in buffering resources can be achieved.
2010, 39(6): 934-939.
doi: 10.3969/j.issn.1001-0548.2010.06.027
Abstract:
A low voltage high efficiency charge pump circuit for passive UHF RFID is presented to increase the operating range of tags. The bias voltage supplied to the main charge pump by the bias circuit is raised by suppressing its load current using a diode connected MOSFET, which greatly reduces the threshold voltage drop in traditional charge pumps, suppresses the reverse leakage current and improves the sensitivity and power conversion efficiency. The charge pump has been fabricated in chartered 0.35 μm CMOS process. Measurement results show that a 275 mV minimum input level is required to generate 1.47 V power supply for 200 kΩ load and efficiency up to 26.2% is achieved. The maximum operating range of the RFID tag with this charge pump is 4.2 m. This design effectively contributes to the good performance of RFID chips.
A low voltage high efficiency charge pump circuit for passive UHF RFID is presented to increase the operating range of tags. The bias voltage supplied to the main charge pump by the bias circuit is raised by suppressing its load current using a diode connected MOSFET, which greatly reduces the threshold voltage drop in traditional charge pumps, suppresses the reverse leakage current and improves the sensitivity and power conversion efficiency. The charge pump has been fabricated in chartered 0.35 μm CMOS process. Measurement results show that a 275 mV minimum input level is required to generate 1.47 V power supply for 200 kΩ load and efficiency up to 26.2% is achieved. The maximum operating range of the RFID tag with this charge pump is 4.2 m. This design effectively contributes to the good performance of RFID chips.
2010, 39(6): 940-943.
doi: 10.3969/j.issn.1001-0548.2010.06.028
Abstract:
A simplified QRD-M algorithm for the IEEE 802.11n STBC-VBLAST is proposed. In order to carry out a limited tree search, each surviving path is expanded only to its partial branches according to the estimation of the symbol to be detected. The proposed scheme can reduce 70% and 90% fundamental operations for 16-QAM and 64-QAM respectively. So the computational complexity is reduced significantly and is more attractive to the VLSI implementation. Simulation results prove that the proposed scheme can achieve a performance very close to the conventional QRD-M algorithm, and yield a tradeoff between the complexity and performance.
A simplified QRD-M algorithm for the IEEE 802.11n STBC-VBLAST is proposed. In order to carry out a limited tree search, each surviving path is expanded only to its partial branches according to the estimation of the symbol to be detected. The proposed scheme can reduce 70% and 90% fundamental operations for 16-QAM and 64-QAM respectively. So the computational complexity is reduced significantly and is more attractive to the VLSI implementation. Simulation results prove that the proposed scheme can achieve a performance very close to the conventional QRD-M algorithm, and yield a tradeoff between the complexity and performance.
2010, 39(6): 944-947.
doi: 10.3969/j.issn.1001-0548.2010.06.029
Abstract:
To avoid the interaction between the cue and the target on the electroencephalogram (EEG) study of inhibition of return (IOR), the low-resolution electromagnetic (LORETA) algorithm is used to localize the source after an exogenous cue in this work. The results show that activations may be divided into three stages. In the early stage, activations are in the prefrontal cortex, the intraparietal cortex, and the contralateral occipito-temporal cortex. In the middle stage, activations are in the frontal cortex and the parietal cortex. In the late stage, activations are in the ipsilateral occipito-parietal cortex. These findings indicate that the generation of IOR includes two neural networks, and support that IOR is related to both attention and motor response inhibition.
To avoid the interaction between the cue and the target on the electroencephalogram (EEG) study of inhibition of return (IOR), the low-resolution electromagnetic (LORETA) algorithm is used to localize the source after an exogenous cue in this work. The results show that activations may be divided into three stages. In the early stage, activations are in the prefrontal cortex, the intraparietal cortex, and the contralateral occipito-temporal cortex. In the middle stage, activations are in the frontal cortex and the parietal cortex. In the late stage, activations are in the ipsilateral occipito-parietal cortex. These findings indicate that the generation of IOR includes two neural networks, and support that IOR is related to both attention and motor response inhibition.
2010, 39(6): 948-951.
doi: 10.3969/j.issn.1001-0548.2010.06.030
Abstract:
Neuronal current magnetic resonance imaging (nc-MRI) can improve the temporal resolution of functional magnetic resonance imaging (fMRI). To explore the feasibility of the nc-MRI, an experiment is designed by, using MR to detect the differences of alpha wave activity under different conditions. One of the conditions is in the darkness with eyes closed to promote alpha wave activity and the other is a visual stimulus with eyes opened to suppress alpha wave activity. Our results show that the MR signal obtained in the alpha band changes significantly, and the changes in the occipital area where the alpha activity is prominent are more consistent. that means, in particular conditions, it is possible to apply nc-MRI to detect neuronal current activity.
Neuronal current magnetic resonance imaging (nc-MRI) can improve the temporal resolution of functional magnetic resonance imaging (fMRI). To explore the feasibility of the nc-MRI, an experiment is designed by, using MR to detect the differences of alpha wave activity under different conditions. One of the conditions is in the darkness with eyes closed to promote alpha wave activity and the other is a visual stimulus with eyes opened to suppress alpha wave activity. Our results show that the MR signal obtained in the alpha band changes significantly, and the changes in the occipital area where the alpha activity is prominent are more consistent. that means, in particular conditions, it is possible to apply nc-MRI to detect neuronal current activity.
2010, 39(6): 952-955,960.
doi: 10.3969/j.issn.1001-0548.2010.06.031
Abstract:
Since the samples for some structural products are very few and the reliability obtained from the physical model has some errors, an integrated reliability function prediction model based on failure data, failure modes and degradation mechanics of the structural product is proposed. First order reliability method (FORM) is implemented to calculate the reliability based on physical model. Both the reliability from failure data and physical model are considered as fuzzy inputs to estimate the parameters in the probability distribution. The proposed reliability function prediction model can accurately predict the time-dependent reliability of the structural product. Reliability-based design and maintenance strategy can be properly made in terms of the predicted time-dependent reliability. An engineering example is used to illustrate the proposed reliability function prediction model.
Since the samples for some structural products are very few and the reliability obtained from the physical model has some errors, an integrated reliability function prediction model based on failure data, failure modes and degradation mechanics of the structural product is proposed. First order reliability method (FORM) is implemented to calculate the reliability based on physical model. Both the reliability from failure data and physical model are considered as fuzzy inputs to estimate the parameters in the probability distribution. The proposed reliability function prediction model can accurately predict the time-dependent reliability of the structural product. Reliability-based design and maintenance strategy can be properly made in terms of the predicted time-dependent reliability. An engineering example is used to illustrate the proposed reliability function prediction model.
2010, 39(6): 956-960.
doi: 10.3969/j.issn.1001-0548.2010.06.032
Abstract:
In order to evaluate the structural reliability under multi-level fatigue loads, an expression of residual strength is derived from the equivalent damage ratio residual strength model and the assumption that the initial strength is a random variable of any distribution. Then a stress-strength interference model under multi-fatigue loads is proposed by employing the universal generating function to calculate the probability characteristics of a function constituted by fatigue loads and residual strength. The fatigue experimental data of 45# steel is used to testify the proposed model. The results indicate that the proposed model can be used to evaluate the structural reliability under multi-level fatigue loads.
In order to evaluate the structural reliability under multi-level fatigue loads, an expression of residual strength is derived from the equivalent damage ratio residual strength model and the assumption that the initial strength is a random variable of any distribution. Then a stress-strength interference model under multi-fatigue loads is proposed by employing the universal generating function to calculate the probability characteristics of a function constituted by fatigue loads and residual strength. The fatigue experimental data of 45# steel is used to testify the proposed model. The results indicate that the proposed model can be used to evaluate the structural reliability under multi-level fatigue loads.
