ISSN 
2023 Vol. 52, No. 2
column
- Special Section on Quantum Information
- Special Section on Bioinformatics
- Special Section for UESTC Youth: Information and Communication Engineering
- Communication and Information Engineering
- Automation Techniques
- Computer Engineering and Applications
- Complexity Sciences
- Bioelectronics
- Mechatronic Engineering
2023, 52(2): 162-167.
doi: 10.12178/1001-0548.2023006
Abstract:
The vulnerability of quantum classifiers under adversarial attacks is one of the fundamental problems in quantum machine learning. The vulnerability of quantum classifiers refers to the property that a quantum classifier may be failed by small perturbations when the quantum system scales up. Such perturbations are also known as quantum adversarial attacks. How to generate small perturbations to fail a quantum classifier is still an open problem. To address this problem, we present a new quantum adversarial attack generation method, the quantum confounding algorithm, which generates perturbations that fail the trained quantum classifier by utilizing the gradient information of the quantum classifier with respect to the input data. Numerical results demonstrate that, compared with the existing quantum adversarial attack generation methods, our quantum confounding algorithm can generate significantly smaller perturbations that lead the quantum classifier to malfunction. This provides a new perspective in understanding the effectiveness and the vulnerability of quantum classifiers.
The vulnerability of quantum classifiers under adversarial attacks is one of the fundamental problems in quantum machine learning. The vulnerability of quantum classifiers refers to the property that a quantum classifier may be failed by small perturbations when the quantum system scales up. Such perturbations are also known as quantum adversarial attacks. How to generate small perturbations to fail a quantum classifier is still an open problem. To address this problem, we present a new quantum adversarial attack generation method, the quantum confounding algorithm, which generates perturbations that fail the trained quantum classifier by utilizing the gradient information of the quantum classifier with respect to the input data. Numerical results demonstrate that, compared with the existing quantum adversarial attack generation methods, our quantum confounding algorithm can generate significantly smaller perturbations that lead the quantum classifier to malfunction. This provides a new perspective in understanding the effectiveness and the vulnerability of quantum classifiers.
2023, 52(2): 168-174.
doi: 10.12178/1001-0548.2022096
Abstract:
The phase matching protocol, which can break through the limitation of key capacity and has been proved by theory and practice, is a kind of twin-field quantum key distribution protocol. Aiming at the adverse effects of the finite data length effect in practical applications, the statistical fluctuation performance of the phase matching protocol is systematically analyzed. Using Gaussian analysis, Chernoff-Hoeffding bounds, and other statistical fluctuation analysis methods, the performances of the three-decoy and two-decoy phase matching protocols under different data lengths were simulated and analyzed combined with linear programming to estimate the relevant parameters. The simulation results show that the phase matching protocol considering statistical fluctuation can still break through the limitation of linear key capacity. When the data length reaches the order of 1016, the key generation rate and the maximum secure transmission distance are both close to ideal values. When the data length is less than or equal to 1013, adding decoy states cannot significantly improve the performance of the phase-matching protocol. As the data length increases, the performance of the system using the Chernoff-Hoeffding bound gradually approaches the performance of the system using the Gaussian analysis method.
The phase matching protocol, which can break through the limitation of key capacity and has been proved by theory and practice, is a kind of twin-field quantum key distribution protocol. Aiming at the adverse effects of the finite data length effect in practical applications, the statistical fluctuation performance of the phase matching protocol is systematically analyzed. Using Gaussian analysis, Chernoff-Hoeffding bounds, and other statistical fluctuation analysis methods, the performances of the three-decoy and two-decoy phase matching protocols under different data lengths were simulated and analyzed combined with linear programming to estimate the relevant parameters. The simulation results show that the phase matching protocol considering statistical fluctuation can still break through the limitation of linear key capacity. When the data length reaches the order of 1016, the key generation rate and the maximum secure transmission distance are both close to ideal values. When the data length is less than or equal to 1013, adding decoy states cannot significantly improve the performance of the phase-matching protocol. As the data length increases, the performance of the system using the Chernoff-Hoeffding bound gradually approaches the performance of the system using the Gaussian analysis method.
2023, 52(2): 175-181.
doi: 10.12178/1001-0548.2022392
Abstract:
Existing acupoint detection (AD) approaches suffer from extra-equipment-dependent, shallow feature representation, and poor accuracy issues. In this work, the AD task is defined as the key-point detection based on visual images by analyzing the task nature. A novel paradigm called facial acupoint detection by reconstruction (FADbR) is designed to achieve the facial AD task. Firstly, the adversarial autoencoder architecture serves as the backbone network based on the self-supervised learning mechanism. The image-to-image reconstruction procedure is performed to enhance the feature representation ability, in which the neural architecture is applied to capture hidden representations and abstract knowledge of the human face. In succession, the FADbR framework is constructed to implement the AD task in a supervised manner by designing the interleaved layers to output the heatmap for each acupoint. Because of the reconstruction procedure, a fine-grained model can be achieved to improve AD performance by the learned facial representations. A new dataset called FAcupoint is built to validate the proposed approach using a public human face dataset. Experimental results on the new dataset demonstrate that the proposed FADbR framework has the ability to extract high-level feature representation to improve AD performance. Most importantly, the FADbR framework can achieve preferred performance with small training samples, which further validates the reconstruction paradigm in this work.
Existing acupoint detection (AD) approaches suffer from extra-equipment-dependent, shallow feature representation, and poor accuracy issues. In this work, the AD task is defined as the key-point detection based on visual images by analyzing the task nature. A novel paradigm called facial acupoint detection by reconstruction (FADbR) is designed to achieve the facial AD task. Firstly, the adversarial autoencoder architecture serves as the backbone network based on the self-supervised learning mechanism. The image-to-image reconstruction procedure is performed to enhance the feature representation ability, in which the neural architecture is applied to capture hidden representations and abstract knowledge of the human face. In succession, the FADbR framework is constructed to implement the AD task in a supervised manner by designing the interleaved layers to output the heatmap for each acupoint. Because of the reconstruction procedure, a fine-grained model can be achieved to improve AD performance by the learned facial representations. A new dataset called FAcupoint is built to validate the proposed approach using a public human face dataset. Experimental results on the new dataset demonstrate that the proposed FADbR framework has the ability to extract high-level feature representation to improve AD performance. Most importantly, the FADbR framework can achieve preferred performance with small training samples, which further validates the reconstruction paradigm in this work.
2023, 52(2): 182-187.
doi: 10.12178/1001-0548.2023002
Abstract:
In recent years, many studies have revealed the relationship between microRNA expression and diseases, especially its close relationship with the occurrence, development and treatment of tumors. However, traditional molecular biology testing methods are time-consuming and expensive, and it is difficult to obtain disease samples. The classifier obtained from imbalanced data set training leads to low accuracy of disease sample recognition. In the face of the above challenges, we propose a new data augmentation algorithm OCF (original data-based conditional generative adversarial network for sample generation) to distinguish health samples from disease samples and mine disease biomarkers, by using conditional generative adversarial networks for data augmentation, followed by feature selection algorithms to reduce the number of features. Finally, the machine learning classifier is used for classification and recognition, and the biomarkers are selected for analysis. The experimental results show that our proposed algorithm has better classification performance, and verify the accuracy of the selected biomarkers.
In recent years, many studies have revealed the relationship between microRNA expression and diseases, especially its close relationship with the occurrence, development and treatment of tumors. However, traditional molecular biology testing methods are time-consuming and expensive, and it is difficult to obtain disease samples. The classifier obtained from imbalanced data set training leads to low accuracy of disease sample recognition. In the face of the above challenges, we propose a new data augmentation algorithm OCF (original data-based conditional generative adversarial network for sample generation) to distinguish health samples from disease samples and mine disease biomarkers, by using conditional generative adversarial networks for data augmentation, followed by feature selection algorithms to reduce the number of features. Finally, the machine learning classifier is used for classification and recognition, and the biomarkers are selected for analysis. The experimental results show that our proposed algorithm has better classification performance, and verify the accuracy of the selected biomarkers.
2023, 52(2): 188-195.
doi: 10.12178/1001-0548.2022163
Abstract:
Aiming at the building layout imaging problem of wide-beam microwave computerized tomography (CT), considering that the electromagnetic waves emitted by antennas have a certain beamwidth in practical applications, the Rytov signal model is established based on the constraint of electromagnetic beam direction graph, and a statistical imaging method of microwave tomography based on the wide-beam signal is proposed. First, by constructing the maximum likelihood function of the imaging matrix of the building layout, the statistical imaging equation is derived. Then, the imaging matrix is updated by the sparse angle-dependent subset data, which realizes the microwave tomography of the building layout under wide-beam signal. The proposed method is suitable for wide-beam microwave CT imaging with sparse sampling angles, and has a strong suppression effect on noise. The validity of the proposed signal model and method are verified by simulation and experimental data.
Aiming at the building layout imaging problem of wide-beam microwave computerized tomography (CT), considering that the electromagnetic waves emitted by antennas have a certain beamwidth in practical applications, the Rytov signal model is established based on the constraint of electromagnetic beam direction graph, and a statistical imaging method of microwave tomography based on the wide-beam signal is proposed. First, by constructing the maximum likelihood function of the imaging matrix of the building layout, the statistical imaging equation is derived. Then, the imaging matrix is updated by the sparse angle-dependent subset data, which realizes the microwave tomography of the building layout under wide-beam signal. The proposed method is suitable for wide-beam microwave CT imaging with sparse sampling angles, and has a strong suppression effect on noise. The validity of the proposed signal model and method are verified by simulation and experimental data.
2023, 52(2): 196-202.
doi: 10.12178/1001-0548.2022359
Abstract:
Fusion of radar and communication (RadCom) is regarded as an important research direction of the next generation of communication technology, which has attracted extensive attention from academia and industry. In this paper, a multiuser orthogonal frequency division multiplexing (OFDM) RadCom scheme based on frequency division multiple access (FDMA) is studied. The key problem lies in the multiuser subcarrier allocation. Different subcarrier allocation methods will lead to different sidelobe levels in range processing, which will affect the ranging performance of radar; when subcarriers are continuously assigned and randomly assigned, there will be obvious sidelobe in range processing of each user, and the performance of radar ranging will be degraded seriously. In order to find a proper subcarrier allocation method that makes the sidelobe level of radar range processing of each user as low as possible, we take the integrated sidelobe level in the range observation window as a cost function. We find that when the number of users is limited and the subcarriers are allocated with equal intervals, the cost function achieves the minimum value, and the sidelobe level of all users is the lowest in the range observation window, so all users can effectively realize the radar detection function. Finally, in the internet of vehicles (IoV) scenario, different subcarrier allocation methods are simulated and compared to verify the effectiveness of the equal interval subcarrier allocation method.
Fusion of radar and communication (RadCom) is regarded as an important research direction of the next generation of communication technology, which has attracted extensive attention from academia and industry. In this paper, a multiuser orthogonal frequency division multiplexing (OFDM) RadCom scheme based on frequency division multiple access (FDMA) is studied. The key problem lies in the multiuser subcarrier allocation. Different subcarrier allocation methods will lead to different sidelobe levels in range processing, which will affect the ranging performance of radar; when subcarriers are continuously assigned and randomly assigned, there will be obvious sidelobe in range processing of each user, and the performance of radar ranging will be degraded seriously. In order to find a proper subcarrier allocation method that makes the sidelobe level of radar range processing of each user as low as possible, we take the integrated sidelobe level in the range observation window as a cost function. We find that when the number of users is limited and the subcarriers are allocated with equal intervals, the cost function achieves the minimum value, and the sidelobe level of all users is the lowest in the range observation window, so all users can effectively realize the radar detection function. Finally, in the internet of vehicles (IoV) scenario, different subcarrier allocation methods are simulated and compared to verify the effectiveness of the equal interval subcarrier allocation method.
2023, 52(2): 203-208.
doi: 10.12178/1001-0548.2022091
Abstract:
With the advantages of high performance, strong robustness, and large service range, unmanned aerial vehicle (UAV) swarm systems have been widely used in military and civil scenarios. UAV swarms need to form specific topology shapes autonomously through topology shaping to achieve efficient swarm collaboration mechanisms for specific mission scenarios. Topology shaping typically involves two aspects: the optimal mapping from the initial topology to the target topology and the optimal topology shaping location. These two aspects affect each other and are directly related to the global energy consumption of the UAV swarm. Based on the joint optimization model of UAV swarm topology shaping with global energy consumption minimization as the goal, a generalized solution framework based on the swarm intelligence algorithm is firstly established, and specific solution methods based on the gray wolf optimizer algorithm (GWO), the equilibrium optimizer algorithm (EO) and the poor and rich optimization algorithm (PRO) are proposed. Then the convergence acceleration strategy for solving the optimization model based on the swarm intelligence algorithm is proposed. Simulation results show the effectiveness of the proposed UAV swarm topology shaping method and indicate that the proposed optimization method can achieve convergence within 8 iterations in a typical topology shaping scenario.
With the advantages of high performance, strong robustness, and large service range, unmanned aerial vehicle (UAV) swarm systems have been widely used in military and civil scenarios. UAV swarms need to form specific topology shapes autonomously through topology shaping to achieve efficient swarm collaboration mechanisms for specific mission scenarios. Topology shaping typically involves two aspects: the optimal mapping from the initial topology to the target topology and the optimal topology shaping location. These two aspects affect each other and are directly related to the global energy consumption of the UAV swarm. Based on the joint optimization model of UAV swarm topology shaping with global energy consumption minimization as the goal, a generalized solution framework based on the swarm intelligence algorithm is firstly established, and specific solution methods based on the gray wolf optimizer algorithm (GWO), the equilibrium optimizer algorithm (EO) and the poor and rich optimization algorithm (PRO) are proposed. Then the convergence acceleration strategy for solving the optimization model based on the swarm intelligence algorithm is proposed. Simulation results show the effectiveness of the proposed UAV swarm topology shaping method and indicate that the proposed optimization method can achieve convergence within 8 iterations in a typical topology shaping scenario.
2023, 52(2): 209-213.
doi: 10.12178/1001-0548.2021208
Abstract:
This paper proposes a stub-loaded gallium nitride field-effect transistor (GaN-FET) phase shifter with ultra-low DC power consumption and high power handling capability. The phase shifter consists of multiple micro-strip stubs with high impedance and GaN-FET loaded. By analyzing the circuit model of the single stub-loaded GaN-FET phase shifter in detail, the formulas for insertion loss and phase shift are obtained. Then, as the basic element, the GaN-FET phase shifter using multiple stubs is developed, which can operate at the predefined frequency band. Due to the employment of the high-Z micro-strip stub, the proposed phase shifter can handle high input power. Following the design theory, the GaN-FET phase shifter using multiple stubs is designed, fabricated and measured. The tested results show that the phase shifter achieves the phase shift of 30° and 60° by controlling the GaN FETs within the frequency range of 9.2 GHz to 9.8 GHz. In addition, the power handling capability is over 10 W, insertion loss less than 1 dB and DC current for control less than 6 μA.
This paper proposes a stub-loaded gallium nitride field-effect transistor (GaN-FET) phase shifter with ultra-low DC power consumption and high power handling capability. The phase shifter consists of multiple micro-strip stubs with high impedance and GaN-FET loaded. By analyzing the circuit model of the single stub-loaded GaN-FET phase shifter in detail, the formulas for insertion loss and phase shift are obtained. Then, as the basic element, the GaN-FET phase shifter using multiple stubs is developed, which can operate at the predefined frequency band. Due to the employment of the high-Z micro-strip stub, the proposed phase shifter can handle high input power. Following the design theory, the GaN-FET phase shifter using multiple stubs is designed, fabricated and measured. The tested results show that the phase shifter achieves the phase shift of 30° and 60° by controlling the GaN FETs within the frequency range of 9.2 GHz to 9.8 GHz. In addition, the power handling capability is over 10 W, insertion loss less than 1 dB and DC current for control less than 6 μA.
2023, 52(2): 214-223.
doi: 10.12178/1001-0548.2022089
Abstract:
In radar communication and electronic warfare, multifunctional integrated load system needs hardware generalization, functions based on software and resource virtualization, so as to realize resource reuse and sharing to the maximum extent. In this paper, we design a polar-reconfigurable antenna array with low profile, ultra-wideband and wide angle scanning, a comprehensive RF microsystem with reconfigurable RF performances, and a high-performance heterogeneous computing platform for flexible resource scheduling and dynamic management. The multifunctional integrated architecture and its key technologies proposed in this paper lay the technical foundation for the function based on software, virtualization and intelligentization of distributed multi-domain intelligent networked electronic systems in the future.
In radar communication and electronic warfare, multifunctional integrated load system needs hardware generalization, functions based on software and resource virtualization, so as to realize resource reuse and sharing to the maximum extent. In this paper, we design a polar-reconfigurable antenna array with low profile, ultra-wideband and wide angle scanning, a comprehensive RF microsystem with reconfigurable RF performances, and a high-performance heterogeneous computing platform for flexible resource scheduling and dynamic management. The multifunctional integrated architecture and its key technologies proposed in this paper lay the technical foundation for the function based on software, virtualization and intelligentization of distributed multi-domain intelligent networked electronic systems in the future.
2023, 52(2): 224-231.
doi: 10.12178/1001-0548.2021388
Abstract:
In this paper, we propose a mobility-aware time synchronization method for multi-gateway LoRa networks. By jointly considering the selection of synchronization timing and spreading factors for both end devices and gateways, the synchronization accuracy is improved and energy consumption is reduced. This paper also proposes a mobility-aware adaptation algorithm to address the impact of end device mobility across multiple gateways. The experiments based on both LoRa testbed and simulations reveal that the proposed method can improve the synchronization accuracy by 60.43% compared with the state-of-the-art works. This paper also discusses the future directions such as transmission scheduling and the legitimacy evolution.
In this paper, we propose a mobility-aware time synchronization method for multi-gateway LoRa networks. By jointly considering the selection of synchronization timing and spreading factors for both end devices and gateways, the synchronization accuracy is improved and energy consumption is reduced. This paper also proposes a mobility-aware adaptation algorithm to address the impact of end device mobility across multiple gateways. The experiments based on both LoRa testbed and simulations reveal that the proposed method can improve the synchronization accuracy by 60.43% compared with the state-of-the-art works. This paper also discusses the future directions such as transmission scheduling and the legitimacy evolution.
2023, 52(2): 232-239.
doi: 10.12178/1001-0548.2021369
Abstract:
Vibration tactile is widely used in the field of human-computer interaction. It plays an important role in direction navigation, graphic display, music expression, and so on. Vibration tactile music interaction is a new direction of vibration tactile interaction. At present, the research on music information represented by vibration and touch is lack of systematic summary. Firstly, this paper systematically discusses the mapping elements of music information represented by vibration and touch, including pitch, melody, timbre, rhythm and loudness. This paper summarizes the applications and interactive design methods of vibration tactile music from three aspects: enhancing music experience, instrument teaching and dance action assistance. Finally, the development prospect of vibration tactile music is prospected, and it is concluded that the applications of tactile music combined with specific scenes will have practical significance and market value.
Vibration tactile is widely used in the field of human-computer interaction. It plays an important role in direction navigation, graphic display, music expression, and so on. Vibration tactile music interaction is a new direction of vibration tactile interaction. At present, the research on music information represented by vibration and touch is lack of systematic summary. Firstly, this paper systematically discusses the mapping elements of music information represented by vibration and touch, including pitch, melody, timbre, rhythm and loudness. This paper summarizes the applications and interactive design methods of vibration tactile music from three aspects: enhancing music experience, instrument teaching and dance action assistance. Finally, the development prospect of vibration tactile music is prospected, and it is concluded that the applications of tactile music combined with specific scenes will have practical significance and market value.
2023, 52(2): 240-246.
doi: 10.12178/1001-0548.2022063
Abstract:
In order to improve the efficiency of transmission line inspection and solve the problem of low accuracy of overhead hanging objects and bird nests detection, a transmission line hidden danger detection algorithm integrating SE attention model into YOLOv5 network is proposed to obtain channel-level global features, improve the sensitivity of the model to channel features, and increase the accuracy of overhead overhang and bird's nest detection. Extensive experiments were conducted on a set of transmission line hazard images, and the results show that the YOLOv5 network with the attention model has an average accuracy of 84.2% for overhead overhangs and 87.4% for bird nests, and the mAP value detected by the proposed method is 2% higher than that directly using the YOLOv5 algorithm.
In order to improve the efficiency of transmission line inspection and solve the problem of low accuracy of overhead hanging objects and bird nests detection, a transmission line hidden danger detection algorithm integrating SE attention model into YOLOv5 network is proposed to obtain channel-level global features, improve the sensitivity of the model to channel features, and increase the accuracy of overhead overhang and bird's nest detection. Extensive experiments were conducted on a set of transmission line hazard images, and the results show that the YOLOv5 network with the attention model has an average accuracy of 84.2% for overhead overhangs and 87.4% for bird nests, and the mAP value detected by the proposed method is 2% higher than that directly using the YOLOv5 algorithm.
2023, 52(2): 247-253.
doi: 10.12178/1001-0548.2022072
Abstract:
his paper proposes a distribution network electrical topology identification algorithm based on a multi-channel adaptive weighted neural network. The algorithm builds a multi-channel 1DCNN (one-dimensional convolutional neural network) model, and uses four types of data: voltage, current, power and power factor, to make the datasets. Feature extraction has been realized through two CNN layers stacked; Meanwhile, an adaptive weighted feature fusion is proposed, it can learn the importance of each channel's feature through neural network adaptively. Datasets collect real consumption data, and multiple sets of experiments are conducted with the number of channels, data types, data dimensions and other parameters. Results show that the proposed algorithm can integrate the advantages of multiple data features, the accuracy of electrical topology identification can reach 99.772%.
his paper proposes a distribution network electrical topology identification algorithm based on a multi-channel adaptive weighted neural network. The algorithm builds a multi-channel 1DCNN (one-dimensional convolutional neural network) model, and uses four types of data: voltage, current, power and power factor, to make the datasets. Feature extraction has been realized through two CNN layers stacked; Meanwhile, an adaptive weighted feature fusion is proposed, it can learn the importance of each channel's feature through neural network adaptively. Datasets collect real consumption data, and multiple sets of experiments are conducted with the number of channels, data types, data dimensions and other parameters. Results show that the proposed algorithm can integrate the advantages of multiple data features, the accuracy of electrical topology identification can reach 99.772%.
2023, 52(2): 254-262.
doi: 10.12178/1001-0548.2022081
Abstract:
Abstract This paper proposes a medical image segmentation algorithm based on object detection. With the help of an object detection network, the precise locations of the object of interest (usually an organ or diseased tissue) in medical images are obtained. And the medical images are cropped according to locations obtained by the object detection network to gain the image patches of objects of interest. By taking the key center area of the image patch of objects of interest as the examplar image and this image patch as the search image, the pixel-wise classification of this image patch is achieved by one Siamese network with the help of a pre-defined threshold. Two datasets (LITS17 and KITS19) are used to evaluate our algorithm. And experiments show that the proposed method can accurately segment organs or diseased tissues in medical images.
Abstract This paper proposes a medical image segmentation algorithm based on object detection. With the help of an object detection network, the precise locations of the object of interest (usually an organ or diseased tissue) in medical images are obtained. And the medical images are cropped according to locations obtained by the object detection network to gain the image patches of objects of interest. By taking the key center area of the image patch of objects of interest as the examplar image and this image patch as the search image, the pixel-wise classification of this image patch is achieved by one Siamese network with the help of a pre-defined threshold. Two datasets (LITS17 and KITS19) are used to evaluate our algorithm. And experiments show that the proposed method can accurately segment organs or diseased tissues in medical images.
2023, 52(2): 263-270.
doi: 10.12178/1001-0548.2022058
Abstract:
Stock market forecasting is a difficult problem in the field of financial analysis. The intrinsic information contained in financial news has a great impact on the stock market performance. In this paper, we propose a BERT-based vector autoregressive network (BVANet), which quantifies financial news sentiment by BERT and then combines it with market performance to construct a financial time series vector autoregressive (VAR) model to achieve stock prediction eventually. The results show that BVANet has improved results in extracting news sentiment information and model prediction compared with traditional algorithms, and the sentiment of news has predictive effect on market performance. This study can provide a practical reference for the application of natural language processing in financial prediction.
Stock market forecasting is a difficult problem in the field of financial analysis. The intrinsic information contained in financial news has a great impact on the stock market performance. In this paper, we propose a BERT-based vector autoregressive network (BVANet), which quantifies financial news sentiment by BERT and then combines it with market performance to construct a financial time series vector autoregressive (VAR) model to achieve stock prediction eventually. The results show that BVANet has improved results in extracting news sentiment information and model prediction compared with traditional algorithms, and the sentiment of news has predictive effect on market performance. This study can provide a practical reference for the application of natural language processing in financial prediction.
2023, 52(2): 271-279.
doi: 10.12178/1001-0548.2022177
Abstract:
A perturbation algorithm with minimum edges is proposed to solve the problem of how to make a specific target path become the shortest path with the minimum cost on the premise of disturbing the minimum edges. The algorithm is based on the shortest path perturbation model with the fewest edges. By introducing the perturbation upper limit constraint of each edge weight, the biobjective mixed integer programming problem of minimizing the number of perturbation edges and minimizing the total cost of perturbation is established. Thus, the shortest path between the nodes of the control network is realized. Compared with the previous minimum cost perturbation algorithm, this method reduces the complexity of perturbation and the risk of disturbance network being detected. Experiments show that the optimal solution reduces the number of perturbed edges by about 27% and has better performance.
A perturbation algorithm with minimum edges is proposed to solve the problem of how to make a specific target path become the shortest path with the minimum cost on the premise of disturbing the minimum edges. The algorithm is based on the shortest path perturbation model with the fewest edges. By introducing the perturbation upper limit constraint of each edge weight, the biobjective mixed integer programming problem of minimizing the number of perturbation edges and minimizing the total cost of perturbation is established. Thus, the shortest path between the nodes of the control network is realized. Compared with the previous minimum cost perturbation algorithm, this method reduces the complexity of perturbation and the risk of disturbance network being detected. Experiments show that the optimal solution reduces the number of perturbed edges by about 27% and has better performance.
2023, 52(2): 280-288.
doi: 10.12178/1001-0548.2022077
Abstract:
In modern information warfare, identifying important targets in the grid can effectively guide the tactics of the attacker or defender. The existing research on grid attack and defense strategy usually ignores the resilient of the real grids, therefore, RHGEle_Rank (important Element identification based on Resilient Heterogeneous Grid) method is proposed to identify important substations and transmission lines in a multi-round attack and defense game of the grid. Based on simulated resilient power network model, a self-recovery strategy for nodes is designed to simulate the self-healing scenario of the substation under multiple rounds of attack and defense game by considering the efficiency of the nodes. Then, according to the heterogeneous characteristics of power network, cascade failure models based on overload and outage modes are constructed. Finally, a greedy algorithm is used to identify the best attack (defense) target in each round of the offense-defense game. Experiments show that power supply capacity of grid can be destroyed to a greater extent by RHGEle_Rank method than that by the traditional degree centrality and betweenness centrality methods. At the same time, "invalid attack" in multi-round attack and defense games can be effectively avoided if network resilience is considered in the important element identification algorithm.
In modern information warfare, identifying important targets in the grid can effectively guide the tactics of the attacker or defender. The existing research on grid attack and defense strategy usually ignores the resilient of the real grids, therefore, RHGEle_Rank (important Element identification based on Resilient Heterogeneous Grid) method is proposed to identify important substations and transmission lines in a multi-round attack and defense game of the grid. Based on simulated resilient power network model, a self-recovery strategy for nodes is designed to simulate the self-healing scenario of the substation under multiple rounds of attack and defense game by considering the efficiency of the nodes. Then, according to the heterogeneous characteristics of power network, cascade failure models based on overload and outage modes are constructed. Finally, a greedy algorithm is used to identify the best attack (defense) target in each round of the offense-defense game. Experiments show that power supply capacity of grid can be destroyed to a greater extent by RHGEle_Rank method than that by the traditional degree centrality and betweenness centrality methods. At the same time, "invalid attack" in multi-round attack and defense games can be effectively avoided if network resilience is considered in the important element identification algorithm.
2023, 52(2): 289-295.
doi: 10.12178/1001-0548.2022366
Abstract:
Emotion regulation and relaxation state assessment aiming at relaxation and stress reduction can help improve the physical and mental health of the people. In order to reduce the difficulty of inducing relaxation caused by the acquisition of physiological signals, this paper uses the attached human sensor to collect the heart rate signal of the subjects to identify the relaxation state. Mixed audio is used to induce the subjects to produce a relaxed mood, and the relaxed state label is converted from the two-dimensional mood scale. The heart rate signal of the subject is collected, and the time domain feature, frequency domain feature and heart rate are extracted from the heart rate signal. Based on multilayer perceptron and long-short-term memory network, a relaxation evaluation model is constructed to realize relaxation state recognition. The experimental results show that, compared with the current research results, the relaxation recognition model proposed in this paper has better classification performance, and may provide a new and reliable method for emotion regulation and relaxation state assessment.
Emotion regulation and relaxation state assessment aiming at relaxation and stress reduction can help improve the physical and mental health of the people. In order to reduce the difficulty of inducing relaxation caused by the acquisition of physiological signals, this paper uses the attached human sensor to collect the heart rate signal of the subjects to identify the relaxation state. Mixed audio is used to induce the subjects to produce a relaxed mood, and the relaxed state label is converted from the two-dimensional mood scale. The heart rate signal of the subject is collected, and the time domain feature, frequency domain feature and heart rate are extracted from the heart rate signal. Based on multilayer perceptron and long-short-term memory network, a relaxation evaluation model is constructed to realize relaxation state recognition. The experimental results show that, compared with the current research results, the relaxation recognition model proposed in this paper has better classification performance, and may provide a new and reliable method for emotion regulation and relaxation state assessment.
2023, 52(2): 296-305.
doi: 10.12178/1001-0548.2022041
Abstract:
Brain imaging is a hot spot field in the biomedical engineering research. Current traditional brain imaging modalities, such as X-ray, magnetic resonance imaging (MRI) and other methods, are mature and have been widely applied for scientific research and clinical diagnosis and treatment comprehensively. However, most of these imaging techniques have limitations such as invasiveness, bulky equipment, and high imaging costs, and are difficult to apply to special populations, such as neonates. This paper introduces several microscopic optical molecular imaging modalities that can be used for brain imaging. Most of these methods are non-invasive, low-cost and excellent in performance, and have broad development prospects. After giving their imaging principles, system components, and key techniques, the existing research results and the current state of these imaging modalities are summarized. Finally, the future development directions of the brain imaging techniques are discussed through enumerating and comparing the merits and limitations among these imaging modalities.
Brain imaging is a hot spot field in the biomedical engineering research. Current traditional brain imaging modalities, such as X-ray, magnetic resonance imaging (MRI) and other methods, are mature and have been widely applied for scientific research and clinical diagnosis and treatment comprehensively. However, most of these imaging techniques have limitations such as invasiveness, bulky equipment, and high imaging costs, and are difficult to apply to special populations, such as neonates. This paper introduces several microscopic optical molecular imaging modalities that can be used for brain imaging. Most of these methods are non-invasive, low-cost and excellent in performance, and have broad development prospects. After giving their imaging principles, system components, and key techniques, the existing research results and the current state of these imaging modalities are summarized. Finally, the future development directions of the brain imaging techniques are discussed through enumerating and comparing the merits and limitations among these imaging modalities.
2023, 52(2): 306-312.
doi: 10.12178/1001-0548.2021299
Abstract:
Wind energy is safe and clean, but, the maintenance of wind turbines which are main device to produce wind energy is a challenge and expensive. Exact values of the performance rates (levels), state transition probability and maintenance cost is unable to get when the wind turbine is modeled as an multi-state system (MSS). Therefore, a maintenance optimization method is proposed for the wind turbine under this epistemic uncertainty. In the proposed method, a selective maintenance model is formulated for the wind turbine generator system, which can be viewed as a multi-state system. The Markov decision process is proposed to resolve the selective maintenance problem. By combining the Markov decision process with the fuzzy theory, the reward and the state distribution of the system are represented as triangular fuzzy numbers to quantify the epistemic uncertainty. By the proposed method, the maintenance decision of the wind turbine generator system can be well-optimized.
Wind energy is safe and clean, but, the maintenance of wind turbines which are main device to produce wind energy is a challenge and expensive. Exact values of the performance rates (levels), state transition probability and maintenance cost is unable to get when the wind turbine is modeled as an multi-state system (MSS). Therefore, a maintenance optimization method is proposed for the wind turbine under this epistemic uncertainty. In the proposed method, a selective maintenance model is formulated for the wind turbine generator system, which can be viewed as a multi-state system. The Markov decision process is proposed to resolve the selective maintenance problem. By combining the Markov decision process with the fuzzy theory, the reward and the state distribution of the system are represented as triangular fuzzy numbers to quantify the epistemic uncertainty. By the proposed method, the maintenance decision of the wind turbine generator system can be well-optimized.
2023, 52(2): 313-320.
doi: 10.12178/1001-0548.2022013
Abstract:
In high speed machining with multi-axis linkage, the dynamic error of machine tool caused by motion parameters is an important factor to produce machining errors. In this paper, the dynamic error and static error models of EtherCAT bus CNC machine tools are established by using the rod instrument QC20-W and the industrial Ethernet probe tool ET2000 as measuring tools. Firstly, based on the homogeneous coordinate transformation theory, the mathematical relationship among the rod length variation and the dynamic and static errors of the machine tool is established. Then, according to the different characteristics of the two kinds of errors, we construct the static error polynomial expressed by coordinate positions and the dynamic error low order sine polynomial expressed by velocity, acceleration angular frequency, and phase. In order to obtain the angular frequency and phase information of velocity and acceleration in sinusoidal polynomial, Wireshark and ET2000 are used to capture the position information of axis, and the velocity and acceleration are obtained by differentiating. Wavelet transform is used to reduce the noise in position information. Finally, the least square theory is used to solve the unknown coefficients in the error expression. The results show that instruction speed is the main factor of dynamic error, and acceleration is the secondary factor. When the acceleration is constant and the instruction speed increases from 20 mm/s to 40 mm/s and 80 mm/s, the proportion of dynamic error in the total error increases from 2.13% to 11.74% and 49.15%. When the command speed is unchanged and the acceleration increases from 200 mm/s2 to 800 mm/s2, the proportion of dynamic error increases from 22.72% to 26.83%, respectively.
In high speed machining with multi-axis linkage, the dynamic error of machine tool caused by motion parameters is an important factor to produce machining errors. In this paper, the dynamic error and static error models of EtherCAT bus CNC machine tools are established by using the rod instrument QC20-W and the industrial Ethernet probe tool ET2000 as measuring tools. Firstly, based on the homogeneous coordinate transformation theory, the mathematical relationship among the rod length variation and the dynamic and static errors of the machine tool is established. Then, according to the different characteristics of the two kinds of errors, we construct the static error polynomial expressed by coordinate positions and the dynamic error low order sine polynomial expressed by velocity, acceleration angular frequency, and phase. In order to obtain the angular frequency and phase information of velocity and acceleration in sinusoidal polynomial, Wireshark and ET2000 are used to capture the position information of axis, and the velocity and acceleration are obtained by differentiating. Wavelet transform is used to reduce the noise in position information. Finally, the least square theory is used to solve the unknown coefficients in the error expression. The results show that instruction speed is the main factor of dynamic error, and acceleration is the secondary factor. When the acceleration is constant and the instruction speed increases from 20 mm/s to 40 mm/s and 80 mm/s, the proportion of dynamic error in the total error increases from 2.13% to 11.74% and 49.15%. When the command speed is unchanged and the acceleration increases from 200 mm/s2 to 800 mm/s2, the proportion of dynamic error increases from 22.72% to 26.83%, respectively.
