ISSN 
2014 Vol. 43, No. 5
2014, 43(5): 641-646.
doi: 10.3969/j.issn.1001-0548.2014.05.001
Abstract:
Due to the restrictions of the airborne radar's volume and weight, the traditional methods are not able to obtain the optimal clutter suppression results under the situation of low complexity. Based on the analysis of the structure of multistage Wiener filter, a multistage Wiener filter with variable diagonal loading is proposed. The backward-recursive algorithm of this method is developed to suppress airborne radar's clutter. Simulation results validate the feasibility and effectiveness of the new method and demonstrate that the proposed method is more robust and outperforms the traditional methods in low dimensional filter structure. Also in airborne radar's clutter suppression, the method can achieve the null-depth of -80dB.
Due to the restrictions of the airborne radar's volume and weight, the traditional methods are not able to obtain the optimal clutter suppression results under the situation of low complexity. Based on the analysis of the structure of multistage Wiener filter, a multistage Wiener filter with variable diagonal loading is proposed. The backward-recursive algorithm of this method is developed to suppress airborne radar's clutter. Simulation results validate the feasibility and effectiveness of the new method and demonstrate that the proposed method is more robust and outperforms the traditional methods in low dimensional filter structure. Also in airborne radar's clutter suppression, the method can achieve the null-depth of -80dB.
2014, 43(5): 647-652.
doi: 10.3969/j.issn.1001-0548.2014.05.002
Abstract:
Based on the analysis of symbol symmetry of orthogonal frequency-division multiplexing (OFDM) modulation and by combining with quadrature up(down)-conversion technique, a half symbol OFDM (HS-OFDM) modulation method based on symbol symmetry is proposed. The BER performance, the spectral efficiency, and the transmission over multipath fading channels are analyzed. By using this method to an OFDM system modulated by binary phase shife keying (BPSK) or pulse amplitude modulation (PAM), the upper limit of the system bandwidth efficiency can be raised to 2 Baud/Hz.
Based on the analysis of symbol symmetry of orthogonal frequency-division multiplexing (OFDM) modulation and by combining with quadrature up(down)-conversion technique, a half symbol OFDM (HS-OFDM) modulation method based on symbol symmetry is proposed. The BER performance, the spectral efficiency, and the transmission over multipath fading channels are analyzed. By using this method to an OFDM system modulated by binary phase shife keying (BPSK) or pulse amplitude modulation (PAM), the upper limit of the system bandwidth efficiency can be raised to 2 Baud/Hz.
2014, 43(5): 653-657.
doi: 10.3969/j.issn.1001-0548.2014.05.003
Abstract:
In this paper, a joint base station and relay station bandwidth allocation algorithm based on greedy search is proposed for IEEE 802.16j network to meet the need of different services and achieve the maximum access users. In the proposed algorithm, the relay station allocates a bandwidth to the unsolicited grant service which has fixed bandwidth, the base station allocates the minimal bandwidth to the other service, and then the greedy search algorithm is utilized to find out the optimal user group and allocate the remaining bandwidth to the users in this group. Simulation results show that the proposed bandwidth allocation algorithm based on greedy search can guarantee the maximum number of admitted users, and the implementation complexity is significantly less than conventional bandwidth allocation algorithms based on exhaustive search.
In this paper, a joint base station and relay station bandwidth allocation algorithm based on greedy search is proposed for IEEE 802.16j network to meet the need of different services and achieve the maximum access users. In the proposed algorithm, the relay station allocates a bandwidth to the unsolicited grant service which has fixed bandwidth, the base station allocates the minimal bandwidth to the other service, and then the greedy search algorithm is utilized to find out the optimal user group and allocate the remaining bandwidth to the users in this group. Simulation results show that the proposed bandwidth allocation algorithm based on greedy search can guarantee the maximum number of admitted users, and the implementation complexity is significantly less than conventional bandwidth allocation algorithms based on exhaustive search.
2014, 43(5): 658-662.
doi: 10.3969/j.issn.1001-0548.2014.05.004
Abstract:
The minimum resolution of the broadband frequency sweeping jamming (BFSJ) bandwidth is set to 1/M of the MFSK signal bandwidth, theoretical derivation of the anti-jamming performance for FH/MFSK systems in noise-free environment and additive white Gaussian noise (AWGN) channels are derived, respectively. The closed-form expressions of bit error rate (BER) are given and validated by computer simulations. The theoretical and simulation results show that the minimum resolution of jamming bandwidth has more influence on the BER performance when the jamming bandwidth becomes smaller. With the increases in the sweeping rate, the BER performance decreases under the given power of sweeping jamming and the bandwidth over AWGN channels.
The minimum resolution of the broadband frequency sweeping jamming (BFSJ) bandwidth is set to 1/M of the MFSK signal bandwidth, theoretical derivation of the anti-jamming performance for FH/MFSK systems in noise-free environment and additive white Gaussian noise (AWGN) channels are derived, respectively. The closed-form expressions of bit error rate (BER) are given and validated by computer simulations. The theoretical and simulation results show that the minimum resolution of jamming bandwidth has more influence on the BER performance when the jamming bandwidth becomes smaller. With the increases in the sweeping rate, the BER performance decreases under the given power of sweeping jamming and the bandwidth over AWGN channels.
2014, 43(5): 663-668.
doi: 10.3969/j.issn.1001-0548.2014.05.005
Abstract:
A spectrum sensing method in the femtocell of long time evolution (LTE) system is proposed. In this method, the femtocells sense the uplink spectrum via energy sensing and discover the active macrocell users through two-steps decision spectrum sensing method. The femtocells identify the traffic types of these active macrocell users and adjust the sensing period according to different traffic types to implement the tradeoffs between throughput and sensing overhead of femtocells. Simulation results indicate that this method can improve the femtocell throughput and system performance when the interference to macrocell users and sensing overhead are effectively controlled.
A spectrum sensing method in the femtocell of long time evolution (LTE) system is proposed. In this method, the femtocells sense the uplink spectrum via energy sensing and discover the active macrocell users through two-steps decision spectrum sensing method. The femtocells identify the traffic types of these active macrocell users and adjust the sensing period according to different traffic types to implement the tradeoffs between throughput and sensing overhead of femtocells. Simulation results indicate that this method can improve the femtocell throughput and system performance when the interference to macrocell users and sensing overhead are effectively controlled.
2014, 43(5): 669-673.
doi: 10.3969/j.issn.1001-0548.2014.05.006
Abstract:
A lower power digital correction model for comparator offset and noise tolerance of successive approximation register (SAR) analog-to-digital converter (ADC) is presented. A fine comparator with smaller offset and noise has penalty of higher power and lower speed. This model involves a faster coarse comparator with less power in the first (n-m-1) cycles to relax those penalties. The errors of the coarse comparator are tolerated by the fine comparator through the redundant comparison cycle and the capacitor at (m+2) cycle. This model is able to tolerate noise and offset errors up to ±2m least significant bit (LSB). A prototype of 10 bit 100 MS/s SAR ADC with this model is simulated in a 0.13μm CMOS technology. The post-simulation results of the prototype layout witnessed an effective number of bits (ENOB) of 9.27 bit are achieved at 100 MS/s with a power consumption of 2.01mW under 1.2V supply, resulting in a figure of merit (FoM) of 33fJ/conv.
A lower power digital correction model for comparator offset and noise tolerance of successive approximation register (SAR) analog-to-digital converter (ADC) is presented. A fine comparator with smaller offset and noise has penalty of higher power and lower speed. This model involves a faster coarse comparator with less power in the first (n-m-1) cycles to relax those penalties. The errors of the coarse comparator are tolerated by the fine comparator through the redundant comparison cycle and the capacitor at (m+2) cycle. This model is able to tolerate noise and offset errors up to ±2m least significant bit (LSB). A prototype of 10 bit 100 MS/s SAR ADC with this model is simulated in a 0.13μm CMOS technology. The post-simulation results of the prototype layout witnessed an effective number of bits (ENOB) of 9.27 bit are achieved at 100 MS/s with a power consumption of 2.01mW under 1.2V supply, resulting in a figure of merit (FoM) of 33fJ/conv.
2014, 43(5): 674-677.
doi: 10.3969/j.issn.1001-0548.2014.05.007
Abstract:
In traditional localization methods based on distance measurement in wireless sensor networks, a lot of optimization algorithms are employed to improve the precision of location. However, the optimization performance may degrade because of the low SNR environment and the non-Gaussian noise distribution. To overcome this problem, the trust region (TR) algorithm based on the conic model is employed for iterative localization. The cone model is used for approximating the objective function of localization, and the optimization process is converted into a series of optimization sub-problems. Simulation results show that the TR based method acquires better performance, especially in low SNR environment.
In traditional localization methods based on distance measurement in wireless sensor networks, a lot of optimization algorithms are employed to improve the precision of location. However, the optimization performance may degrade because of the low SNR environment and the non-Gaussian noise distribution. To overcome this problem, the trust region (TR) algorithm based on the conic model is employed for iterative localization. The cone model is used for approximating the objective function of localization, and the optimization process is converted into a series of optimization sub-problems. Simulation results show that the TR based method acquires better performance, especially in low SNR environment.
2014, 43(5): 678-684.
doi: 10.3969/j.issn.1001-0548.2014.05.008
Abstract:
The channel estimation model suitable for underwater acoustic OFDM communication signals is studied in this paper. Based on the analysis of existing channel estimation algorithms, the matching pursuit (MP) algorithm was improved by introducing a threshold adjusting mechanism which can effectively avoid false paths. An iterative correction method is used to decrease the estimation error caused by non-orthogonality. The improved MP algorithm can achieve a comparative estimation precision with OMP, but has smaller computation amount. In order to overcome the dependences on the information of channel's length, a new channel estimation algorithm is proposed to adjust pre-estimated channel length automatically based on the real-time evaluation of channel estimation's quality. Besides, the relations between the number of pilots and channel estimation effects are studied and a self-adjusting OFDM underwater acoustic communication scheme is presented for achieving reliable communication with lesser bandwidth recourse losses. The correctness and validity of the presented results are validated by simulations and experiments.
The channel estimation model suitable for underwater acoustic OFDM communication signals is studied in this paper. Based on the analysis of existing channel estimation algorithms, the matching pursuit (MP) algorithm was improved by introducing a threshold adjusting mechanism which can effectively avoid false paths. An iterative correction method is used to decrease the estimation error caused by non-orthogonality. The improved MP algorithm can achieve a comparative estimation precision with OMP, but has smaller computation amount. In order to overcome the dependences on the information of channel's length, a new channel estimation algorithm is proposed to adjust pre-estimated channel length automatically based on the real-time evaluation of channel estimation's quality. Besides, the relations between the number of pilots and channel estimation effects are studied and a self-adjusting OFDM underwater acoustic communication scheme is presented for achieving reliable communication with lesser bandwidth recourse losses. The correctness and validity of the presented results are validated by simulations and experiments.
2014, 43(5): 685-690.
doi: 10.3969/j.issn.1001-0548.2014.05.009
Abstract:
Increasingly complicated very large scale integration (VLSI) design and system-on-chip (SoC) makes Field Programmable Gate Arrays (FPGA)-based emulation necessary. As a design is downloaded into a FPGA-based emulator, invisible internal nodes of the design pose a challenge for design debugging. To address the issue, a RTL-level runtime debugging system is proposed. The user can not only select sample signals, triggering signals and statements in RTL codes, but can change triggering mode or sample window runtime as well. The proposed debugging system supports three debugging modes including embedded logic analyzer (ELA) mode, scan mode and snapshot mode. All debugging modes use a unified external interface in order to make the debugging system suitable for both Altera and Xilinx FPGAs. Experiment results show that the ELA mode consumes nearly the same resources while the scan mode and snapshot mode consumes fewer resources, compared with SignalTap and ChipScope.
Increasingly complicated very large scale integration (VLSI) design and system-on-chip (SoC) makes Field Programmable Gate Arrays (FPGA)-based emulation necessary. As a design is downloaded into a FPGA-based emulator, invisible internal nodes of the design pose a challenge for design debugging. To address the issue, a RTL-level runtime debugging system is proposed. The user can not only select sample signals, triggering signals and statements in RTL codes, but can change triggering mode or sample window runtime as well. The proposed debugging system supports three debugging modes including embedded logic analyzer (ELA) mode, scan mode and snapshot mode. All debugging modes use a unified external interface in order to make the debugging system suitable for both Altera and Xilinx FPGAs. Experiment results show that the ELA mode consumes nearly the same resources while the scan mode and snapshot mode consumes fewer resources, compared with SignalTap and ChipScope.
2014, 43(5): 691-695.
doi: 10.3969/j.issn.1001-0548.2014.05.010
Abstract:
A directional antenna of four directions is designed based on the broadband miniaturization technology of distribution medium load and concentrated resistance load. The gain of antenna main lobe is 5.4 dB and in the center frequency the gain of anterior lobe is 28 dB larger than the back lobe. The ratio of the peak-to-peak value of forward time-domain radiation pulse to backward time-domain radiation pulse is greater than 3. The trailing of time domain radiation pulse waveform is less than one fifth of main peak amplitude. The joint simulation of both omni-directional and directional borehole radar antennas shows that the directionality of the designed antenna is improved. The antenna is fabricated to make the test and the measured results of the antenna's S11 are consistent with the simulation results.
A directional antenna of four directions is designed based on the broadband miniaturization technology of distribution medium load and concentrated resistance load. The gain of antenna main lobe is 5.4 dB and in the center frequency the gain of anterior lobe is 28 dB larger than the back lobe. The ratio of the peak-to-peak value of forward time-domain radiation pulse to backward time-domain radiation pulse is greater than 3. The trailing of time domain radiation pulse waveform is less than one fifth of main peak amplitude. The joint simulation of both omni-directional and directional borehole radar antennas shows that the directionality of the designed antenna is improved. The antenna is fabricated to make the test and the measured results of the antenna's S11 are consistent with the simulation results.
2014, 43(5): 696-699.
doi: 10.3969/j.issn.1001-0548.2014.05.011
Abstract:
A novel photonic-band-gap (PBG) horn antenna model is presented by using a metallic PBG structure which is formed by square arrays of metal posts. The lowest global band gap of the two-dimensional PBG structure is determined numerically and the gap is used to reflect and focus the antenna radiation. The horn antenna is designed by using the photonic crystals and simulated by using electromagnetic simulation software. The operating frequency of THz antenna is 300GHz and the gain of THz antenna gain is 17.6dB. Compared with the traditional antenna, the gain is increased, main lobe narrowed, antical lobe weakened, and the directivity improved.
A novel photonic-band-gap (PBG) horn antenna model is presented by using a metallic PBG structure which is formed by square arrays of metal posts. The lowest global band gap of the two-dimensional PBG structure is determined numerically and the gap is used to reflect and focus the antenna radiation. The horn antenna is designed by using the photonic crystals and simulated by using electromagnetic simulation software. The operating frequency of THz antenna is 300GHz and the gain of THz antenna gain is 17.6dB. Compared with the traditional antenna, the gain is increased, main lobe narrowed, antical lobe weakened, and the directivity improved.
2014, 43(5): 700-705.
doi: 10.3969/j.issn.1001-0548.2014.05.012
Abstract:
This paper proposes an extended state observer (ESO) based control strategy to effectively reject the coupling, disturbance and uncertainties in the boiler-turbine unit. The parameter tuning rule of ESO and its effect on estimation and noise suppression ability are analysed. A 300 MW coal-fired boiler-turbine unit model is simulated. The results of the set point tracking simulation as well as the perturbation simulation illustrate that with well-tuned ESO parameters, the proposed control strategy achieves satisfactory control results and strong disturbance rejection ability. With ESO, the disturbance and system uncertainties can be estimated as an extended state and then compensated by the control law. Besides, the ESO based controller has simple structure, thus brings a lot convenience in engineering applications.
This paper proposes an extended state observer (ESO) based control strategy to effectively reject the coupling, disturbance and uncertainties in the boiler-turbine unit. The parameter tuning rule of ESO and its effect on estimation and noise suppression ability are analysed. A 300 MW coal-fired boiler-turbine unit model is simulated. The results of the set point tracking simulation as well as the perturbation simulation illustrate that with well-tuned ESO parameters, the proposed control strategy achieves satisfactory control results and strong disturbance rejection ability. With ESO, the disturbance and system uncertainties can be estimated as an extended state and then compensated by the control law. Besides, the ESO based controller has simple structure, thus brings a lot convenience in engineering applications.
2014, 43(5): 706-711.
doi: 10.3969/j.issn.1001-0548.2014.05.013
Abstract:
The short circuit current calculation is an important foundation for planning and protection of distribution grid with distributed generation. The transient characteristics of induction generator (IG) are investigated when the unsymmetrical fault occurs in the distribution network, and the analytic formula of its stator current under unsymmetrical short circuit is derived. The impact of an IG rotor speed and the fault initial time (or phase) to the short circuit current under unsymmetrical fault are studied when the short circuit occurs. According to the equation of rotor motion and its steady-state circuits in the positive and negative sequence component networks, an evaluation method to determine the current peak of unsymmetrical short circuit for an IG in distribution grid is proposed by considering its rotor speed variations. The correctness of the method is verified by using the fifth electromagnetic transient model of IG in PSCAD/EMTDC simulation software.
The short circuit current calculation is an important foundation for planning and protection of distribution grid with distributed generation. The transient characteristics of induction generator (IG) are investigated when the unsymmetrical fault occurs in the distribution network, and the analytic formula of its stator current under unsymmetrical short circuit is derived. The impact of an IG rotor speed and the fault initial time (or phase) to the short circuit current under unsymmetrical fault are studied when the short circuit occurs. According to the equation of rotor motion and its steady-state circuits in the positive and negative sequence component networks, an evaluation method to determine the current peak of unsymmetrical short circuit for an IG in distribution grid is proposed by considering its rotor speed variations. The correctness of the method is verified by using the fifth electromagnetic transient model of IG in PSCAD/EMTDC simulation software.
2014, 43(5): 712-716.
doi: 10.3969/j.issn.1001-0548.2014.05.014
Abstract:
There is no method for model checking real-time systems within the same real-time interval logic. To this end, we restrict a real-time logic, called Timed Interval Temporal Logic (TITL), on discrete time domain. And then, we use a TITL formula to construct an interval model and another TITL formula to describe an interval property. On the basis of this, we formalize a novel approach for model checking within the same logical framework based on TITL. The validity and correctness of the method are proved at last.
There is no method for model checking real-time systems within the same real-time interval logic. To this end, we restrict a real-time logic, called Timed Interval Temporal Logic (TITL), on discrete time domain. And then, we use a TITL formula to construct an interval model and another TITL formula to describe an interval property. On the basis of this, we formalize a novel approach for model checking within the same logical framework based on TITL. The validity and correctness of the method are proved at last.
2014, 43(5): 717-723.
doi: 10.3969/j.issn.1001-0548.2014.05.015
Abstract:
The classical single kernelized rough set model ignores the interaction between different kernelized relations. In order to solve this problem, this paper is devoted to the construction of the fuzzy rough set model based on multi-kernelized granulation. In this study, the optimistic and pessimistic rough set model, which is derived from a family of the kernelized relations, is deeply explored to multi-kernelized granulation space by defining the S-T multi-kernelized lower and upper approximation operators. Finally, we apply these measures to evaluate and select features of classification problems. The experimental results verify the interaction in different granulating relations.
The classical single kernelized rough set model ignores the interaction between different kernelized relations. In order to solve this problem, this paper is devoted to the construction of the fuzzy rough set model based on multi-kernelized granulation. In this study, the optimistic and pessimistic rough set model, which is derived from a family of the kernelized relations, is deeply explored to multi-kernelized granulation space by defining the S-T multi-kernelized lower and upper approximation operators. Finally, we apply these measures to evaluate and select features of classification problems. The experimental results verify the interaction in different granulating relations.
2014, 43(5): 724-729.
doi: 10.3969/j.issn.1001-0548.2014.05.016
Abstract:
Real networks such as e-mail, co-author and peer-to-peer networks can be modeled as graphs. Community mining on graphs has attracted more and more attentions in recent years. It not only can help to identify the overall structures of networks, but also can help to discover the latent rules of community evolution. Community mining on dynamic graphs has not been studied thoroughly, although that on static graphs has been exploited extensively. Based on time-sequence, the community mining including community detection and analysis on dynamic graphs is researched in this paper. And a two-step model is presented to discover the dynamic community structure. The effectiveness and efficiency of the model are validated by experiments on real networks. Results show that the model has a good trade-off between the effectiveness and efficiency in discovering communities.
Real networks such as e-mail, co-author and peer-to-peer networks can be modeled as graphs. Community mining on graphs has attracted more and more attentions in recent years. It not only can help to identify the overall structures of networks, but also can help to discover the latent rules of community evolution. Community mining on dynamic graphs has not been studied thoroughly, although that on static graphs has been exploited extensively. Based on time-sequence, the community mining including community detection and analysis on dynamic graphs is researched in this paper. And a two-step model is presented to discover the dynamic community structure. The effectiveness and efficiency of the model are validated by experiments on real networks. Results show that the model has a good trade-off between the effectiveness and efficiency in discovering communities.
2014, 43(5): 730-735.
doi: 10.3969/j.issn.1001-0548.2014.05.017
Abstract:
This paper studies the transitional state in the software defined networking (SDN) networks during the SDN network update process. A route optimization scheme is proposed after the brief review of the logic inconsistency issues in SDN networks due to the delay in the network. In this scheme, the timing sequence of the control rules is tuned adaptively on the control plane so as to ensure that the control logic in the whole network is consistent. In addition, an objective optimization model based on the average network latency of nodes is constructed. With this model, we find the minimum-delay route while retaining the logic consistency in the SDN network. The simulation results demonstrate that the proposed scheme is feasible and effective.
This paper studies the transitional state in the software defined networking (SDN) networks during the SDN network update process. A route optimization scheme is proposed after the brief review of the logic inconsistency issues in SDN networks due to the delay in the network. In this scheme, the timing sequence of the control rules is tuned adaptively on the control plane so as to ensure that the control logic in the whole network is consistent. In addition, an objective optimization model based on the average network latency of nodes is constructed. With this model, we find the minimum-delay route while retaining the logic consistency in the SDN network. The simulation results demonstrate that the proposed scheme is feasible and effective.
2014, 43(5): 736-741.
doi: 10.3969/j.issn.1001-0548.2014.05.018
Abstract:
In order to solve the problems that complex conditional statements increase the complexity of program, and affect the understandability, testability, maintainability, and extendibility of existing code, a novel approach for automatic refactoring of complex conditional statements based on abstract syntax tree and polymorphism is proposed. The approach can be used to detect the refactoring opportunities for complex conditional statements and implement automatic refactoring. In this approach, the source code file is transformed to an abstract syntax tree at first; then all of the conditional statements are detected and the candidate statements which meet the preconditions are identified; and finally, the conditional statements are refactored automatically by introducing polymorphism, and each branch of the them is encapsulated into a subclass. Four projects are tested to identify refactoring opportunities and implement automatic refactoring. The results show that the precision of the identification algorithm for refactoring opportunities is 100%. Meanwhile, the approach can execute automatic refactoring correctly. The regression testing shows that none error is imported after refactoring. In addition, this approach has good efficiency and the execution time has a linear relationship with the size of system. It means that this approach can be used for projects of different scales.
In order to solve the problems that complex conditional statements increase the complexity of program, and affect the understandability, testability, maintainability, and extendibility of existing code, a novel approach for automatic refactoring of complex conditional statements based on abstract syntax tree and polymorphism is proposed. The approach can be used to detect the refactoring opportunities for complex conditional statements and implement automatic refactoring. In this approach, the source code file is transformed to an abstract syntax tree at first; then all of the conditional statements are detected and the candidate statements which meet the preconditions are identified; and finally, the conditional statements are refactored automatically by introducing polymorphism, and each branch of the them is encapsulated into a subclass. Four projects are tested to identify refactoring opportunities and implement automatic refactoring. The results show that the precision of the identification algorithm for refactoring opportunities is 100%. Meanwhile, the approach can execute automatic refactoring correctly. The regression testing shows that none error is imported after refactoring. In addition, this approach has good efficiency and the execution time has a linear relationship with the size of system. It means that this approach can be used for projects of different scales.
2014, 43(5): 742-748.
doi: 10.3969/j.issn.1001-0548.2014.05.019
Abstract:
There are several difficulties in using a partitional clustering algorithm to deal with image segmentation problem including choosing the correct number of clusters without any prior knowledge, measuring the image datasets with complicated manifold structures and reducing the computation time. In this paper, an automatic immune clonal clustering method using manifold distance is applied to image segmentation. This method can automatically determine the number of clusters, measure the complicated manifold dataset by using manifold distance, and less computation time by using super-pixels instead of pixels. Experimental results on four artificial data sets and four Berkeley images show that the novel method outperforms the k-means algorithm and the GCUK algorithm.
There are several difficulties in using a partitional clustering algorithm to deal with image segmentation problem including choosing the correct number of clusters without any prior knowledge, measuring the image datasets with complicated manifold structures and reducing the computation time. In this paper, an automatic immune clonal clustering method using manifold distance is applied to image segmentation. This method can automatically determine the number of clusters, measure the complicated manifold dataset by using manifold distance, and less computation time by using super-pixels instead of pixels. Experimental results on four artificial data sets and four Berkeley images show that the novel method outperforms the k-means algorithm and the GCUK algorithm.
2014, 43(5): 749-753.
doi: 10.3969/j.issn.1001-0548.2014.05.020
Abstract:
Computer vision algorithms for individual tasks such as object recognition, detection and segmentation have shown impressive results in the recent years. The next challenge is to integrate all these algorithms and address the problem of scene understanding. A new higher order conditional random field (CRF) model is proposed to get semantic segmentation and object detection simultaneously. Specifically, the proposed higher order CRF model consists of low-order potentials and improved detector potentials. To avoid wrong recognition caused by the confidence given by the initial detector, the first-and-second-order pooling and logistic regression are adopted to improve the detector potential. Experimental results show that the proposed model achieves significant improvement over the baseline methods on MSRC 21-class and PASCAL VOC 2007 datasets.
Computer vision algorithms for individual tasks such as object recognition, detection and segmentation have shown impressive results in the recent years. The next challenge is to integrate all these algorithms and address the problem of scene understanding. A new higher order conditional random field (CRF) model is proposed to get semantic segmentation and object detection simultaneously. Specifically, the proposed higher order CRF model consists of low-order potentials and improved detector potentials. To avoid wrong recognition caused by the confidence given by the initial detector, the first-and-second-order pooling and logistic regression are adopted to improve the detector potential. Experimental results show that the proposed model achieves significant improvement over the baseline methods on MSRC 21-class and PASCAL VOC 2007 datasets.
2014, 43(5): 754-757.
doi: 10.3969/j.issn.1001-0548.2014.05.021
Abstract:
Support vector machine (SVM) is widely used in wind speed forecasting and the forecasted results are verified well. With the applications get more intensive, there exist two problems in SVM. One is it is too sensitive to noises and the other is it can not fully use the information included in the samples. In view of this, a fuzzy manifold-based support vector machine (FMSVM) is proposed in this paper to solve the above problems and further improve the generalization capability of SVM. FMSVM introduces the fuzzy techniques to decrease the influence of the noises. Meanwhile, FMSVM takes boundary data between classes, data distributions and manifold seriously. The comparative experiments show that FMSVM performs better than SVM on the wind datasets of a certain wind farm.
Support vector machine (SVM) is widely used in wind speed forecasting and the forecasted results are verified well. With the applications get more intensive, there exist two problems in SVM. One is it is too sensitive to noises and the other is it can not fully use the information included in the samples. In view of this, a fuzzy manifold-based support vector machine (FMSVM) is proposed in this paper to solve the above problems and further improve the generalization capability of SVM. FMSVM introduces the fuzzy techniques to decrease the influence of the noises. Meanwhile, FMSVM takes boundary data between classes, data distributions and manifold seriously. The comparative experiments show that FMSVM performs better than SVM on the wind datasets of a certain wind farm.
2014, 43(5): 758-763.
doi: 10.3969/j.issn.1001-0548.2014.05.022
Abstract:
An attribute extraction method based on weakly supervised learning is proposed in the paper. The training corpus is automatically acquired from natural language texts by using structured attribute information from knowledgebase. To solve the problem that noise exists in the training corpus, an optimization method based on keywords filtering is proposed. N-pattern features extraction method is proposed which can relieve to some extent the data sparsity problem of traditional n-gram features. Experiment data are downloaded from Hudong Baike. Structured attribute information is extracted from infoboxes of Hudong Baike and used to construct knowledgebase. Training data and testing data are acquired from encyclopedia entry texts. Experiment results show that the method of keywords filtering can effectively improve the quality of training corpus, and achieve better performance of attribute extraction by using n-pattern features, compared with traditional n-gram features.
An attribute extraction method based on weakly supervised learning is proposed in the paper. The training corpus is automatically acquired from natural language texts by using structured attribute information from knowledgebase. To solve the problem that noise exists in the training corpus, an optimization method based on keywords filtering is proposed. N-pattern features extraction method is proposed which can relieve to some extent the data sparsity problem of traditional n-gram features. Experiment data are downloaded from Hudong Baike. Structured attribute information is extracted from infoboxes of Hudong Baike and used to construct knowledgebase. Training data and testing data are acquired from encyclopedia entry texts. Experiment results show that the method of keywords filtering can effectively improve the quality of training corpus, and achieve better performance of attribute extraction by using n-pattern features, compared with traditional n-gram features.
2014, 43(5): 764-768.
doi: 10.3969/j.issn.1001-0548.2014.05.023
Abstract:
In order to improve the robustness of binary tree pruning based topology inference algorithm, a topology inference algorithm based on nonparametric hypothesis test is proposed. In this method, the binary tree is obtained by using classical binary tree inference algorithm; wilcoxon rank sum test method is applied to test which internal nodes should be removed; and finally all the internal nodes which should be removed are deleted to generate the real topology. Simulation results show that the algorithm can get higher topology inference accuracy than the method based on threshold comparison.
In order to improve the robustness of binary tree pruning based topology inference algorithm, a topology inference algorithm based on nonparametric hypothesis test is proposed. In this method, the binary tree is obtained by using classical binary tree inference algorithm; wilcoxon rank sum test method is applied to test which internal nodes should be removed; and finally all the internal nodes which should be removed are deleted to generate the real topology. Simulation results show that the algorithm can get higher topology inference accuracy than the method based on threshold comparison.
2014, 43(5): 769-774.
Abstract:
The fast rerouting paths of resilient routing layers are not short enough to deliver traffic efficiently. To solve this problem, a resilient routing sub-layer building algorithm based on Partheno-genetic algorithm is proposed. The matrix expression of resilient routing layers is given. The optimizing models respectively with average minimum rerouting path length and weighted minimum rerouting path length are put forward. The method and steps to solve the optimizing problem based on Partheno-genetic algorithm are given. Experimental results show that the optimal topology structure of resilient routing layers could be acquired by the optimizing algorithm. Compared with minimum algorithm and Rich algorithm, the optimizing algorithm can reduce the rerouting path length and improve the efficiency of traffic forwarding.
The fast rerouting paths of resilient routing layers are not short enough to deliver traffic efficiently. To solve this problem, a resilient routing sub-layer building algorithm based on Partheno-genetic algorithm is proposed. The matrix expression of resilient routing layers is given. The optimizing models respectively with average minimum rerouting path length and weighted minimum rerouting path length are put forward. The method and steps to solve the optimizing problem based on Partheno-genetic algorithm are given. Experimental results show that the optimal topology structure of resilient routing layers could be acquired by the optimizing algorithm. Compared with minimum algorithm and Rich algorithm, the optimizing algorithm can reduce the rerouting path length and improve the efficiency of traffic forwarding.
2014, 43(5): 775-780.
doi: 10.3969/j.issn.1001-0548.2014.05.025
Abstract:
Based on the characteristics of RIFD middleware, the paper adds an intrusion detection system in RFID middleware to gather the feature vector from the comparison between intrusion data stream with RIFD internal data stream and establish finite automation according to intrusion data steam corresponding to sub-module. The efficiency of the intrusion detection system is improved via the consolidation of some of finite automation. Besides, the proposed system can achieve the classification of RFID attacks from finite automation, and even detect some unknown attacks through analyzing the nature of the attacks. By randomizing the order of access events which has mixed attack events in fixed proportion, the system can achieve stable detection rates in 3 rounds tests.
Based on the characteristics of RIFD middleware, the paper adds an intrusion detection system in RFID middleware to gather the feature vector from the comparison between intrusion data stream with RIFD internal data stream and establish finite automation according to intrusion data steam corresponding to sub-module. The efficiency of the intrusion detection system is improved via the consolidation of some of finite automation. Besides, the proposed system can achieve the classification of RFID attacks from finite automation, and even detect some unknown attacks through analyzing the nature of the attacks. By randomizing the order of access events which has mixed attack events in fixed proportion, the system can achieve stable detection rates in 3 rounds tests.
2014, 43(5): 781-786.
doi: 10.3969/j.issn.1001-0548.2014.05.026
Abstract:
Secure multi-party computation (SMC) is a multi-party cooperative computation conducted by a group of participants on the premise that they do not trust each other and will not disclose any of their private information. Secure intersection computation (SIC), an important sub-concern of SMC, is focused on how to acquire intersection through cooperative computing and ensure the security of privacy, which has a significant application prospect in regard to social networks, military and commercial fields. In view of the low efficiency and complexity of current intersection computation methods, a high-efficient and secure intersection computation protocol is proposed. Under such a protocol, each element within a set will be converted into a dot in planar space and the intersection can be computed and acquired through the distance relations between these dots. At last, the correctness, security and complexity of the protocol are analyzed and verified through simulation experiment.
Secure multi-party computation (SMC) is a multi-party cooperative computation conducted by a group of participants on the premise that they do not trust each other and will not disclose any of their private information. Secure intersection computation (SIC), an important sub-concern of SMC, is focused on how to acquire intersection through cooperative computing and ensure the security of privacy, which has a significant application prospect in regard to social networks, military and commercial fields. In view of the low efficiency and complexity of current intersection computation methods, a high-efficient and secure intersection computation protocol is proposed. Under such a protocol, each element within a set will be converted into a dot in planar space and the intersection can be computed and acquired through the distance relations between these dots. At last, the correctness, security and complexity of the protocol are analyzed and verified through simulation experiment.
2014, 43(5): 787-793.
doi: 10.3969/j.issn.1001-0548.2014.05.027
Abstract:
Part accuracy in high-speed NC machining is often difficult to control, especial for large structure aviation. Small changes in some dynamic factors may lead to errors exceeding the allowable value. A method integrated machine on dynamic error factors and static geometric errors is presented to build the surface data of actual workpiece. Each servo axis movement is simulated and composed by kinematics of the joint bodies of the machine for calculating the actual position and attitude of the tool. The actual milling point is solved by the envelope theory and the final part of the surface is obtained by the surface forming method. A case study is analyzed through the "S" specimen sample, which is verified the composition of surface methods and get the milling error in influence of dynamic factors, such as the gap, the position loop and speed loop. This method can estimate the milling accuracy on large parts in advance, and help to adjust the dynamic parameters of the complex servo system. It provides a basis for better precision control effect and economic costs.
Part accuracy in high-speed NC machining is often difficult to control, especial for large structure aviation. Small changes in some dynamic factors may lead to errors exceeding the allowable value. A method integrated machine on dynamic error factors and static geometric errors is presented to build the surface data of actual workpiece. Each servo axis movement is simulated and composed by kinematics of the joint bodies of the machine for calculating the actual position and attitude of the tool. The actual milling point is solved by the envelope theory and the final part of the surface is obtained by the surface forming method. A case study is analyzed through the "S" specimen sample, which is verified the composition of surface methods and get the milling error in influence of dynamic factors, such as the gap, the position loop and speed loop. This method can estimate the milling accuracy on large parts in advance, and help to adjust the dynamic parameters of the complex servo system. It provides a basis for better precision control effect and economic costs.
2014, 43(5): 794-800.
doi: 10.3969/j.issn.1001-0548.2014.05.028
Abstract:
Aiming at the non-linearity as well as the uncertainty of parameters and external interference in pneumatic hydraulic combination control system (PHCC), the function approximation technique (FAT) based the multiple-surface sliding control (MSSC) algorithm is presented in this paper. The Lyapunov method is applied to acquire the stability of closed-loop system and realize the robust impedance control of PHCC system. The impedance control simulation researches are conducted under the 7mm/s repeatedly slope and 0.2Hz sine signal respectively. The simulation results show that the control strategy is able to perform good compliant control with good robustness and the system performs well in both free-space and compliant tracking movements.
Aiming at the non-linearity as well as the uncertainty of parameters and external interference in pneumatic hydraulic combination control system (PHCC), the function approximation technique (FAT) based the multiple-surface sliding control (MSSC) algorithm is presented in this paper. The Lyapunov method is applied to acquire the stability of closed-loop system and realize the robust impedance control of PHCC system. The impedance control simulation researches are conducted under the 7mm/s repeatedly slope and 0.2Hz sine signal respectively. The simulation results show that the control strategy is able to perform good compliant control with good robustness and the system performs well in both free-space and compliant tracking movements.
