2015 Vol. 45 No. 1
column
Abstract：
Based on the dynamic BFGS (Broyden-Fletcher-Goldfarb-Shanno) method, an uncalibrated visual servoing control approach was presented for the real time tracking of a moving targets. The method directly estimated, global Hessian matrix containing the residual (the Hessian matrix of an object function), thus reducing the cost of computation. Meanwhile, it solved the singularity problem of the Hessian matrix. According to the approximate affine model of the mapping relations between the joint variables map and the image plane, an estimation of the image Jacobian matrix was obtained, which increased the robustness in tracking a dynamic target. Based on the Matlab Robotools, a mechanical arm visual tracking system with three degrees of freedom was constructed. By means of simulation experiments, the presented method was compared with the direct computing method of the residual term and the D-DFP method. The results verify the good tracking performance of the D-BFGS method.
Based on the dynamic BFGS (Broyden-Fletcher-Goldfarb-Shanno) method, an uncalibrated visual servoing control approach was presented for the real time tracking of a moving targets. The method directly estimated, global Hessian matrix containing the residual (the Hessian matrix of an object function), thus reducing the cost of computation. Meanwhile, it solved the singularity problem of the Hessian matrix. According to the approximate affine model of the mapping relations between the joint variables map and the image plane, an estimation of the image Jacobian matrix was obtained, which increased the robustness in tracking a dynamic target. Based on the Matlab Robotools, a mechanical arm visual tracking system with three degrees of freedom was constructed. By means of simulation experiments, the presented method was compared with the direct computing method of the residual term and the D-DFP method. The results verify the good tracking performance of the D-BFGS method.
Abstract：
Due to the effect of outside meteorological conditions, greenhouse covering materials, greenhouse structure and the growth and variety of greenhouse crops and their cultivation methods, a greenhouse temperature system has the characteristics of large time delay, nonlinearity, strong external noise disturbances, time variance. Parameter modeling can hardly describe model structures online. A method was thus proposed, which uses the finite impulse response(FIR) model to describe the temperature system and identify the time delay through the sparsity of FIR sequences. First, the sparsity of FIR sequences were analyzed. Then, according to the compressed sensing theory, a relatively small amount of data to recover the FIR sequences by solving the sparse optimization problems, hereby obtaining the time delay property of the system. Finally, the parameters of FIR model were identified. The time delay of the outside temperature, outside solar radiation, cooling pad, is 6 minutes, 1 minute and 1 minute, respectively. These results are consistent with the mechanism model of the greenhouse temperature system. As the control equipment is incapable of continuous control, the “on” and “off” status of the equipment was brought into the model which was built under the effect of the Wet Curtain-Fan. The fitting of the model was 9468%, 9414% when the Wet Curtain-Fan was on or off, suggesting that the model has higher credibility.
Due to the effect of outside meteorological conditions, greenhouse covering materials, greenhouse structure and the growth and variety of greenhouse crops and their cultivation methods, a greenhouse temperature system has the characteristics of large time delay, nonlinearity, strong external noise disturbances, time variance. Parameter modeling can hardly describe model structures online. A method was thus proposed, which uses the finite impulse response(FIR) model to describe the temperature system and identify the time delay through the sparsity of FIR sequences. First, the sparsity of FIR sequences were analyzed. Then, according to the compressed sensing theory, a relatively small amount of data to recover the FIR sequences by solving the sparse optimization problems, hereby obtaining the time delay property of the system. Finally, the parameters of FIR model were identified. The time delay of the outside temperature, outside solar radiation, cooling pad, is 6 minutes, 1 minute and 1 minute, respectively. These results are consistent with the mechanism model of the greenhouse temperature system. As the control equipment is incapable of continuous control, the “on” and “off” status of the equipment was brought into the model which was built under the effect of the Wet Curtain-Fan. The fitting of the model was 9468%, 9414% when the Wet Curtain-Fan was on or off, suggesting that the model has higher credibility.
Abstract：
Considering the effect of gyro output noise in a velocity loop control system on a gyro-stabilized platform, the adaptive strong tracking Kalman filter with iterative estimation of system parameter perturbation was designed. Combined with the model reference adaptive control (MRAC) system, the disturbance isolation performance of the system with the Kalman filter designed was studied by means of comparative simulation experiments with PI control system. The results show that the adaptive strong tracking Kalman filter proposed may further improve the isolation performance of the disturbance control system, and that especially in the case of half compensated non-linear friction as unmodeled uncertainty, not only can the designed filter work stably, but the isolation performance of MRAC system and PI control system can also be improved significantly.
Considering the effect of gyro output noise in a velocity loop control system on a gyro-stabilized platform, the adaptive strong tracking Kalman filter with iterative estimation of system parameter perturbation was designed. Combined with the model reference adaptive control (MRAC) system, the disturbance isolation performance of the system with the Kalman filter designed was studied by means of comparative simulation experiments with PI control system. The results show that the adaptive strong tracking Kalman filter proposed may further improve the isolation performance of the disturbance control system, and that especially in the case of half compensated non-linear friction as unmodeled uncertainty, not only can the designed filter work stably, but the isolation performance of MRAC system and PI control system can also be improved significantly.
Abstract：
Faraday medium was introduced into circular polarized erbium-doped fiber ring laser gyroscope to break the frequency-locked phenomenon. Using Jones matrix, the polarization of Eigen polarization of the laser was proposed. Compared with the laser without Faraday medium, the polarization has not changed. Meanwhile, the parsing expression of the output signal of the gyro was deduced, which indicated that the unlocked fiber ring laser gyroscope can be realized due to the introduction of the Faraday effect.
Faraday medium was introduced into circular polarized erbium-doped fiber ring laser gyroscope to break the frequency-locked phenomenon. Using Jones matrix, the polarization of Eigen polarization of the laser was proposed. Compared with the laser without Faraday medium, the polarization has not changed. Meanwhile, the parsing expression of the output signal of the gyro was deduced, which indicated that the unlocked fiber ring laser gyroscope can be realized due to the introduction of the Faraday effect.
Abstract：
A new method was proposed for designing two-channel filter banks (FBs) with causal-stable IIR filters. By using IIR filters with a cosine-rolloff transition band, the flatness condition required for two-channel NPR FB was automatically satisfied. Instead of designing the frequency magnitude responses of the analysis filters, the power spectra of the desired filters were designed by solving a quasi-convex problem. When the solution was found, the analysis filters desired can be obtained by spectral factorization. The polyphase components of the analysis filters were assumed to have an identical denominator to simplify the PR condition. Two-channel NPR IIR FB so obtained has a reasonably low reconstruction error and can be employed as the initial guess to constrained nonlinear optimization software for designing the PR IIR FB.
A new method was proposed for designing two-channel filter banks (FBs) with causal-stable IIR filters. By using IIR filters with a cosine-rolloff transition band, the flatness condition required for two-channel NPR FB was automatically satisfied. Instead of designing the frequency magnitude responses of the analysis filters, the power spectra of the desired filters were designed by solving a quasi-convex problem. When the solution was found, the analysis filters desired can be obtained by spectral factorization. The polyphase components of the analysis filters were assumed to have an identical denominator to simplify the PR condition. Two-channel NPR IIR FB so obtained has a reasonably low reconstruction error and can be employed as the initial guess to constrained nonlinear optimization software for designing the PR IIR FB.
Abstract：
The definitions of the Lee complete ρ weight enumerator and the exact complete ρ weight enumerator over Mn×s(Rk)(u2i=0,uiuj=ujui) were given, and the MacWilliams identities with respect to the RT metric for these two weight enumerators of linear codes over Mn×s(Rk) were obtained, respectively. Finally, two examples were presented to illustrate our obtained results.
The definitions of the Lee complete ρ weight enumerator and the exact complete ρ weight enumerator over Mn×s(Rk)(u2i=0,uiuj=ujui) were given, and the MacWilliams identities with respect to the RT metric for these two weight enumerators of linear codes over Mn×s(Rk) were obtained, respectively. Finally, two examples were presented to illustrate our obtained results.
Abstract：
In view of the factorization of (xn－1) in F2［x］, the minimal generating set and rank of (1+u)-constacyclic codes with an arbitrary length over the ring R=F2+uF2+u2F2 were studied. A new Gray map from R to F42 was defined, the structures and generator polynomials of the Gray image of a linear (1+u)-constacyclic code with an arbitrary length were determined, and some optimal binary linear cyclic codes were obtained.
In view of the factorization of (xn－1) in F2［x］, the minimal generating set and rank of (1+u)-constacyclic codes with an arbitrary length over the ring R=F2+uF2+u2F2 were studied. A new Gray map from R to F42 was defined, the structures and generator polynomials of the Gray image of a linear (1+u)-constacyclic code with an arbitrary length were determined, and some optimal binary linear cyclic codes were obtained.
Abstract：
An adaptive Hough transform (AHT) method was proposed, which aims at reducing effects of the quantization unit of the parameter space on Hough transform(HT) in detecting line features. First, the sample model was built up by using samples and computing parameters of the model. Then, according to changes in the model parameters and sample distributions,the method was established to get the appropriate quantization parameters. Finally, the optimized quantization units were obtained and applied to feature extraction in a structured environment. The results show that the proposed method can optimize the quantization units, reduce the line detection error,and improve detection accuracy.
An adaptive Hough transform (AHT) method was proposed, which aims at reducing effects of the quantization unit of the parameter space on Hough transform(HT) in detecting line features. First, the sample model was built up by using samples and computing parameters of the model. Then, according to changes in the model parameters and sample distributions,the method was established to get the appropriate quantization parameters. Finally, the optimized quantization units were obtained and applied to feature extraction in a structured environment. The results show that the proposed method can optimize the quantization units, reduce the line detection error,and improve detection accuracy.
Abstract：
Improving measure accuracy and system security of nuclear instruments is of great importance. In general, improving accuracy will cause a decrease in system security and response speed. So a new sliding-window accumulation algorithm was given, which could improve measurement accuracy in radiant intensity detection, and the system security and response speed remain largely unchanged at the same time. A detailed algorithm analysis based on Monte Carlo method was presented, which indicated the scope of application of the algorithm.
Improving measure accuracy and system security of nuclear instruments is of great importance. In general, improving accuracy will cause a decrease in system security and response speed. So a new sliding-window accumulation algorithm was given, which could improve measurement accuracy in radiant intensity detection, and the system security and response speed remain largely unchanged at the same time. A detailed algorithm analysis based on Monte Carlo method was presented, which indicated the scope of application of the algorithm.
Abstract：
Traditional machine learning methods have lower classification performance when dealing with class imbalanced data. A hierarchical classification model for class imbalanced data was thus proposed. With an AdaBoost classifier as its basis classifier, the model builds mathematical models by the features and false positive rates of the classifier, and demonstrates that parameters of the hierarchical classification model could be calculated. First, the hierarchical classification tree was as the structure, and then the classification cost of the hierarchical classification tree mode was obtained as well as a quantitative and mathematical description of the features of each layer. Finally, the classification cost could be converted to a optimization problem, and the solving process of the optimization problem was given. Meanwhile, results of the hierarchical classification are presented. Experiments have been conducted on UCI dataset, and the results show that the proposed method has higher AUC and F-measure compared to many existing class-imbalanced learning methods.
Traditional machine learning methods have lower classification performance when dealing with class imbalanced data. A hierarchical classification model for class imbalanced data was thus proposed. With an AdaBoost classifier as its basis classifier, the model builds mathematical models by the features and false positive rates of the classifier, and demonstrates that parameters of the hierarchical classification model could be calculated. First, the hierarchical classification tree was as the structure, and then the classification cost of the hierarchical classification tree mode was obtained as well as a quantitative and mathematical description of the features of each layer. Finally, the classification cost could be converted to a optimization problem, and the solving process of the optimization problem was given. Meanwhile, results of the hierarchical classification are presented. Experiments have been conducted on UCI dataset, and the results show that the proposed method has higher AUC and F-measure compared to many existing class-imbalanced learning methods.
Abstract：
By introducing a waiting queue, an original performance evaluation model for a single server based on the finite capacity M/M/1/N-PS model was constructed. For this model, two exact expressions were derived for both users average service response time and average waiting time through solving the differential equation of system. Then the steady-state mean waiting time threshold and mean service time threshold were presented. Also, to provide a scientific quantitative reference for the design of the server, an evaluation method relative to users was proposed to estimate the performance of a server. This evaluation method is based on the principle that server service quality should be evaluated by the users. At last, users average waiting time and mean response time were calculated in this evaluation model with the number of concurrent users N=1 and N→∞, respectively. These results were then compared with those of the typical M/M/1 queuing model and the infinite capacity M/M/1-PS model, respectively.
By introducing a waiting queue, an original performance evaluation model for a single server based on the finite capacity M/M/1/N-PS model was constructed. For this model, two exact expressions were derived for both users average service response time and average waiting time through solving the differential equation of system. Then the steady-state mean waiting time threshold and mean service time threshold were presented. Also, to provide a scientific quantitative reference for the design of the server, an evaluation method relative to users was proposed to estimate the performance of a server. This evaluation method is based on the principle that server service quality should be evaluated by the users. At last, users average waiting time and mean response time were calculated in this evaluation model with the number of concurrent users N=1 and N→∞, respectively. These results were then compared with those of the typical M/M/1 queuing model and the infinite capacity M/M/1-PS model, respectively.
Abstract：
In order to solve the problem that aircraft arrival scheduling and sequencing(ASS) has difficulty when meeting changes of the aircraft messages in dynamic environments, an optimization model based on receding horizon control(RHC) was proposed for the dynamic ASS problems, and optimized sequence of the aircraft in a horizon was saved as a heuristic message for ASS in the next horizon. Then an RHC-based genetic algorithm (GA) with multiple local searches (RHC-MLSGA) was designed to solve the model, and an initialization strategy for population was given on the basis of the saved optimization message. Due to the the fact that existing GA may easily fall into local peak and that GA with single local search can not obtain remarkable performance in convergence and satisfactory solution, different local searches were employed at different stages in the proposed RHC-MLSGA, among which directed local search adjusts the individual maximum searching speed according to gene structures and fitness of the individual and benchmark individual. A large number of experiments show the validity of the proposed model and algorithm and the stability of the algorithm when solving ASS problems in dynamic environments. Several conclusions about the characteristics of ASS problems have been drawn from results of the experiments as well.
In order to solve the problem that aircraft arrival scheduling and sequencing(ASS) has difficulty when meeting changes of the aircraft messages in dynamic environments, an optimization model based on receding horizon control(RHC) was proposed for the dynamic ASS problems, and optimized sequence of the aircraft in a horizon was saved as a heuristic message for ASS in the next horizon. Then an RHC-based genetic algorithm (GA) with multiple local searches (RHC-MLSGA) was designed to solve the model, and an initialization strategy for population was given on the basis of the saved optimization message. Due to the the fact that existing GA may easily fall into local peak and that GA with single local search can not obtain remarkable performance in convergence and satisfactory solution, different local searches were employed at different stages in the proposed RHC-MLSGA, among which directed local search adjusts the individual maximum searching speed according to gene structures and fitness of the individual and benchmark individual. A large number of experiments show the validity of the proposed model and algorithm and the stability of the algorithm when solving ASS problems in dynamic environments. Several conclusions about the characteristics of ASS problems have been drawn from results of the experiments as well.
