Computer and Modernization

Fault Diagnosis of Pumping Unit Based on 1D-CNN-LSTM Attention Network

WANG Lei, ZHANG Xiao-dong, DAI Huan

2023, 0(04): 1-6.

Asbtract ( 305 )

PDF (1482KB) ( 72 )

References | Related Articles | Metrics

Aiming at the problems of complex feature extraction， large amount of model parameters and low diagnostic efficiency in traditional fault diagnosis methods of pumping unit based on dynamometer diagram， this paper proposes a fault diagnosis method based on 1D-CNN-LSTM attention network. The dynamometer diagram is converted into a load displacement sequence as the network input， the one-dimensional convolutional neural network （1D-CNN） is used to extract local features of the sequence while reducing sequence length. Considering the temporal characteristics of the sequence， the long-short-term memory （LSTM） network is further used to extract temporal features of the sequence. In order to highlight the impact of key features， the attention mechanism is introduced to give higher attention weights to temporal features related to fault type. Finally， the weighted features are input into a fully connected layer， and the Softmax classifier is used to realize fault diagnosis. The experimental results show that the average accuracy， precision， recall and F1 value of the proposed method reach 99.13%， 99.35%， 99.17% and 99.25%， respectively， and the model size is only 98 kB. Compared with other methods based on feature engineering， it has higher diagnostic accuracy and generalization. Compared with other methods based on two-dimensional convolutional neural network （2D-CNN） model， it significantly reduces model parameters and training time， improves the efficiency of fault diagnosis.

Automatic Classification Method of CNC Machine Tool Fault Text Based on CNN-BiLSTM

XU Ya-xin, HE Ze-en, XU Xu-kan

2023, 0(04): 7-14.

Asbtract ( 181 )

PDF (1685KB) ( 52 )

References | Related Articles | Metrics

Small and medium-sized CNC machine tool firms have accumulated a large amount of fault maintenance data recorded in manual text during operation and maintenance. In order to accomplish efficient and accurate classification and help maintenance personnel carry out their work efficiently， this paper proposes a fault text classification and prediction approach based on convolution neural network and bi-directional long-short-term memory network. Firstly， the pre-processing is completed by creating a professional feature word database， and Word2Vec is used to train the word vector. Secondly， after the CNN layer extracts local features from the text vector， context features are extracted from the forward and backward LSTM. After the feature fusion and weighting of CNN and BiLSTM layers in the full connected layer， the full connected layer finds the output with the highest probability as the prediction result through the Softmax activation function， and presents the prediction accuracy of each category with the confusion matrix. Based on the fault data of an enterprise in the Yangtze River Delta， this paper makes an experimental analysis， and compares it with a single CNN and BiLSTM model. The experimental results indicate that the prediction accuracy of the new method is up to 94%， the average accuracy is increased by 11 percentage points， and the P value， R value and F value are all up to 95%， which can be used as an effective method in the field of small data volume fault text classification.

Path Planning of Fire Robot Based on Improved Bidirectional A* Algorithm

DU Chuan-sheng, GAO Huan-bing, HOU Yu-xiang, WANG Zi-jian

2023, 0(04): 15-19.

Asbtract ( 123 )

PDF (1552KB) ( 25 )

References | Related Articles | Metrics

In order to solve the global path planning problem of fire robot in complex closed environment， a synchronous bidirectional A* algorithm based on improved heuristic function is proposed. Firstly， the traditional one-way search of A* is changed to synchronous two-way search， and the forward and reverse search target nodes are dynamically defined. Secondly， a “normalization” factor function is added to the evaluation function to prevent the disjoint of forward and reverse search paths， and a dynamic weight function is added to the estimated cost function to reduce the generation of redundant nodes in the search process. In order to solve the problem of many inflection points and unsmooth search path， a solution of corner optimization and local smoothing based on Bezier curve is proposed. Finally， through simulation comparison and real environment experiments， the advantages of the improved algorithm in path length， search time， number of traversing nodes and inflection points are proved， and the effectiveness of the algorithm is verified.

Multi-robot Path Planning Based on Double Fuzzy Inference and Improved DWA Algorithm

WANG Zi-wei

2023, 0(04): 20-25.

Asbtract ( 226 )

PDF (2420KB) ( 52 )

References | Related Articles | Metrics

Aiming at the difficulties—balance the performance indexes such as reachability and safety with time in unknown complex scene—faced by existing multi-robot system path planning methods， an improved DWA （dynamic window approach） algorithm based on two-layer fuzzy inference is proposed. First， the linear velocity fuzzy controller and the steering angle fuzzy controller output the base pose to ensure the flexibility and safety of the robot path-planning process. Then， comparing with the traditional DWA algorithm， the obstacle distance evaluation function is improved and the danger zone-related evaluation function is also incorporated to achieve multi-robot collision avoidance. Also， the robustness and global performance are improved by extending the evaluation function and the weight parameters. Finally， the two-layer fuzzy inference is fused with the improved DWA algorithm， so the two-layer fuzzy controller is used to determine the approximate speed and direction， based on which the precise speed and steering angle are output using the improved DWA. Simulation experiments show that the proposed algorithm generates smoother trajectories and improves the operational efficiency and safety of multi-robot path planning.

Event Extraction Method Based on BERT-BiLSTM-Attention Hybrid Model

WEI Xin, HE Xiao-hai, TENG Qi-zhi, QING Lin-bo, CHEN Hong-gang

2023, 0(04): 26-31.

Asbtract ( 330 )

PDF (1263KB) ( 75 )

References | Related Articles | Metrics

Event extraction is one of the basic tasks in the information extraction’s field， which is aims to extract structured information from unstructured text. The majority of the existing event extraction methods which are based on machine reading comprehension model directly detect and classify the input text trigger words， and to some extent ignore the prediction error caused by judging whether the input text is an event. Therefore， this paper proposes an event extraction method based on BERT-BiLSTM-Attention hybrid model. This method takes BERT-based machine reading comprehension model as the basic model， adopts multi-round question-and-answer method， and adds event classification detection module on the basis of existing machine reading comprehension model to reduce prediction error. BiLSTM model is combined with attention mechanism to form historical session information module to more effectively filter out important information and integrate it into a reading comprehension model. The event extraction experiments are conducted on ACE2005， and the results show that the accuracy， recall and F1 value are improved by 7.8 percentage points， 4.6 percentage points and 5.4 percentage points， respectively， compared with the basic model， which has certain advantages.

Cross-project Software Defect Number Prediction Method Based on Stacked

Denoising Autoencoders

2023, 0(04): 32-38.

Asbtract ( 95 )

PDF (1845KB) ( 32 )

References | Related Articles | Metrics

In the application of software defect prediction technology， the project to be predicted may be a brand new project， or the historical data of the project to be predicted is insufficient. One solution is to use a project （source project） with sufficient data to build a model to complete the prediction of a new project （target project）， and mainly use traditional machine learning methods to perform feature transfer learning on the source project and the target project to complete defect prediction. There is a large difference in the distribution of data between different projects， and the feature representation ability learned by traditional machine methods is weak and the defect prediction performance is poor. In response to this problem， a cross-item defect prediction method based on stacked denoising autoencoders is proposed from the perspective of deep learning. This method combines stacked denoising autoencoders and maximum mean difference distance， which can effectively extract the transferable deep-level feature representation of source items and target items， based on which an effective defect number prediction model can be trained. The experimental results show that compared with the classical cross-item defect prediction methods Burak filtering method， Peters filtering method， TCA and TCA+ on Relink dataset and AEEEM dataset， this method achieves the best prediction results in most cases.

Multi-objective Optimization of Aggregate Production Planning in Iron-steel Plant Based on ESI Cooperation Model

ZHANG Qi-qi, CHEN Qun

2023, 0(04): 39-46.

Asbtract ( 81 )

PDF (1141KB) ( 28 )

References | Related Articles | Metrics

Based on early supplier involvement cooperation mode， the aggregate production planning problem of iron-steel plant （core enterprises） with multi-products， multi-process， multi-downstream enterprises in the supply chain environment is researched. In order to maximize the capacity utilization rate of core enterprises and the sales profit rate of downstream enterprises conjointly， a multi-objective optimization model is established. At the same time， the process restriction of the production capacity of each process on different products， the balance principle of production and sales， the demand expectation of downstream enterprises and the profit of further processing and sales of products are considered. Based on the Pareto dominant theory， a multi-objective multi-population cooperative updating particle swarm optimization algorithm （MMCPSO） is proposed to solve the problem， the external Pareto dominant solution is used to guide the updating of the particles in the population， and the heuristic repair rules for the infeasible solutions are designed， and the unfeasible solutions are heuristically repaired by using the sales profit ranking of product from the downstream enterprises， so as to improve the algorithm performance. The simulation results show that the algorithm is feasible and effective.

Pseudo-color Enhancement Method for Quantum Images Based on IBM Qiskit

LIU Zhi-fei, ZHU Shang-chao, WEI Zhan-hong, ZANG Yi-ming, SUN Wen-tao, HU Guan-shi

2023, 0(04): 47-55.

Asbtract ( 200 )

PDF (4493KB) ( 69 )

References | Related Articles | Metrics

Aiming at the problem of quantum image enhancement， a pseudo-color enhancement method for quantum images based on rainbow coding is proposed. Firstly， the NEQR （Novel Enhanced Quantum Representation） model is used to represent grayscale images， then the quantum circuit of the RGB three-channel color conversion module is designed and optimized， and finally the QRMW （Quantum Representation of Multi Wavelength Images） model is used to represent pseudo-color images. In order to verify the effectiveness of the proposed method， 2×2 and 32×32 NEQR grayscale images are prepared on the IBM quantum computing framework Qiskit， and QRMW pseudo-color images of corresponding sizes are generated by measuring the collapse of the quantum circuit. The experimental results show that， compared with the classical and existing quantum image pseudo-color enhancement methods， this method only requires 958 quantum fundamental gates when processing images with a size of 2n×2n and a color depth of 2q. The time complexity is constant-level O（1）， and the space complexity is O（2n+2q+3）， which significantly reduces the quantum cost， and the information entropy and sharpness indicators of the processed image are good.

YU Peng-fei， LI Hao， HE Xiu-feng， HONG Zhen-hua， LIU Yu-chen

YU Peng-fei, LI Hao, HE Xiu-feng, HONG Zhen-hua, LIU Yu-chen

2023, 0(04): 56-61.

Asbtract ( 97 )

PDF (5430KB) ( 39 )

References | Related Articles | Metrics

Aiming at the problem that the number of homonymous feature points is insufficient or even mismatched due to the large-scale difference between images in image matching， this paper proposes a heterologous image SIFT matching method （BS-SIFT） using word bag model to estimate the scale difference. By sensing the scale difference between the images to be matched in advance， this method transforms the heterogenous image matching into starting at the same scale， improves the interior point rate of matching， and then increases the number of matching points of large-scale difference images. Firstly， by aggregating continuously changing image feature points of different scales in the feature space and reallocating image features of each scale to feature center， the feature distribution relationship of each scale is obtained. Secondly， the scale descriptor between images to be matched is obtained by combining the spatial information entropic weighting of image feature center. Finally， the best image scale difference can be obtained by analyzing the distance distribution of scale descriptors. The experimental results show that the BS-SIFT algorithm proposed in this paper can still achieve good results in image matching with a scale difference of more than 10 times. Compared with the classical SIFT algorithm， the algorithm proposed in this paper can significantly obtain more homonymous feature points while achieving higher efficiency， and the matching accuracy is improved by at least 9 percentage points and up to 37 percentage points.

An Improved MOJAYA/D Algorithm for Image Segmentation

LIU Hui, ZOU Feng, CHEN De-bao, JI Xu-ying, ZHANG Yan

2023, 0(04): 62-72.

Asbtract ( 109 )

PDF (3123KB) ( 42 )

References | Related Articles | Metrics

To deal with the multi-objective optimization problem， the multi-objective JAYA algorithm based on decomposition is proposed. Based on the decomposition-based multi-objective algorithm， the JAYA algorithm is extended to the multi-objective optimization field. Meanwhile， a Lévy flight strategy is introduced to enhance the perturbation of the algorithm， and a feedback learning phase is added to improve the individual learning ability， resulting in the improvement of the algorithm’s diversity and the ability of global optimization search. To verify the performance of the proposed algorithm， it is compared with several classical multi-objective algorithms on the ZDT and DTLZ test functions. The results show that MOJAYA/D outperforms the other algorithms in both convergence and diversity. Finally， the proposed algorithm is applied to the image segmentation problem under multiple objective criteria. The segmentation results show that MOJAYA/D is very effective in dealing with image segmentation problems.

An RGB-D Indoor Scene Classification Method Based on Improved Convolutional Neural Network

ZHU Yuan-ye, NI Jian-jun, TANG Guang-yi

2023, 0(04): 73-77.

Asbtract ( 135 )

PDF (1948KB) ( 33 )

References | Related Articles | Metrics

RGB-D indoor scene classification is a challenging task. In this field， convolutional neural network has yielded excellent outcomes in terms of scene classification. However， many problems arise in the immediate application of traditional convolutional neural networks to indoor scene classification due to the multiple objectives， complex layout of indoor scenes， and the similarity existed between different categories of scenes. Aiming at these problems， an improved RGB-D indoor scene classification method based on convolutional neural networks is proposed， including two branches， one of which is a global feature extraction branch based on ResNet-18 and the other is a fusion branch of depth and semantic information. The features obtained from the two branches are fused for the purpose of indoor scene classification. Experimental results based on the SUN RGB-D dataset have proven the superiority of the proposed method in contrast to existing comparison methods.

Atrioventricular Plane Displacement Detection and Reconstruction Based on CMR Images

ZHU Zhuo-yue, HUANG Huan, SONG Qing-ling, HOU Ju-pan

2023, 0(04): 78-82.

Asbtract ( 75 )

PDF (2031KB) ( 28 )

References | Related Articles | Metrics

The traditional methods of detecting atrioventricular plane displacement have the disadvantages of incomplete measurement information and inaccurate tracking of atrioventricular plane feature points， resulting in distortion of the atrioventricular plane displacement （AVPD） curve. Cardiovascular magnetic resonance （CMR） images can be used to detect room plane displacement more accurately. First， a discriminative correlation filter with channel and spatial reliability （CSR-DCF） is used to enhance the tracking ability of feature points in the atrioventricular plane. Secondly， a three-dimensional atrioventricular plane polyhedron is constructed based on the spatial information of the cardiac MRI images to evaluate the atrioventricular plane displacement as a whole. Finally， the displacement curve of the atrioventricular plane is reconstructed by principal component analysis （PCA）. Experiments show that the reconstructed atrioventricular plane displacement retains more than 96% of the original data information， and the atrioventricular plane displacement curve is smoother and conforms to physiological characteristics.

Lightweight Speech Emotion Recognition for Data Enhancement

CUI Chen-lu, CUI Lin,

2023, 0(04): 83-89.

Asbtract ( 144 )

PDF (2978KB) ( 39 )

References | Related Articles | Metrics

The use of deep learning for speech emotion recognition requires a large amount of training data. In this paper， the original speech is enhanced by adding Gaussian white noise and shifting the waveform to generate new speech signals in the preprocessing stage， which not only improves the recognition accuracy but also enhances the robustness of the model， given the shortage of existing speech emotion databases and the defects of overfitting caused by the small amount of data. At the same time， due to the excessive amount of parameters of the common convolutional neural network， a lightweight model is proposed， which consists of separable convolutional and gated recurrent units. Firstly， MFCC features are extracted from the original speech as the input of the model， and secondly， separable convolutional is used to extract the spatial information of speech， and gated recurrent units extract the temporal information of speech so that the temporal and spatial information can be used to characterize the speech emotion at the same time. It can make the prediction results more accurate. Finally， a fully connected layer with softmax is fed to complete the sentiment classification. The experimental results show that the model in this paper can not only obtain higher accuracy but also compress the model by about 50% compared with the benchmark model.

Entity Multi-expression Method for 3D Simulation of Space

LIU Cong, NIE Yun, WANG Guo-wei

2023, 0(04): 90-94.

Asbtract ( 84 )

PDF (1424KB) ( 28 )

References | Related Articles | Metrics

With the deepening of manned spaceflight and deep space exploration programs， the complexity and reliability requirements of space missions are increasing. The 3D simulation system in aerospace missions can provide an intuitive and convenient simulation analysis method for the exploration and engineering design of new aerospace technologies， and plays an important role in improving the reliability of aerospace missions. The continuous in-depth development of aerospace missions puts forward higher requirements for 3D simulation. The situation expression technology for space entities in the existing 3D simulation technologies has problems such as lack of intuition and accuracy in the context of aerospace tasks. In this paper， an adaptive entity representation method is designed based on the characteristics and requirements of situational representation in aerospace missions. The method dynamically adjusts the representation of entity objects in the 3D scene by analyzing the entity importance level， so as to improve the overall expression ability of the 3D scene， and then improve the multi-expression problem of space entities in aerospace missions.

Nonlinear Trajectory Tracking Control of Quadrotor UAV

YANG Zong-yue, SHI Zheng-hua

2023, 0(04): 95-100.

Asbtract ( 105 )

PDF (1929KB) ( 38 )

References | Related Articles | Metrics

Aiming at the problem of system instability in the case of quadrotor trajectory tracking at high frequency， a trajectory tracking method based on dynamic inverse （DI） technology is designed. Since the quadrotor is a typical under actuated system， this paper considers the double closed-loop design， and verifies the tracking performance of the controller at low frequency and high frequency through various numerical experiments， and compares it with the PID controller. The simulation results show that when the frequency is 0.6π rad/s， the mean square error of position tracking of the proposed method is 0.0514 m less than that of the PID method. When the frequency is greater than 1.15π rad/s， the PID control method makes the system unstable， but the proposed method can still achieve better trajectory tracking.

Intelligent Traffic Light Control System Based on Machine Vision

GUO Zhi-dong, LIU Jun, YANG Fan, ZHOU Xin, CHEN Liang-liang

2023, 0(04): 101-105.

Asbtract ( 169 )

PDF (1392KB) ( 41 )

References | Related Articles | Metrics

In order to complete the real-time traffic light command task accurately， flexibly and efficiently， machine vision is introduced to design a set of intelligent traffic light control system based on machine vision to meet the needs of intelligent urban road traffic lights at this stage. The system uses OV2640 camera to take pictures of the road conditions， collect image information， and store the collected data. Then， based on the series of algorithms YOLO （You Only Look Once） in visual processing， it carries out real-time recognition of the road conditions in the data， and transmits the recognition results to the timing module. Finally， it obtains the most timely data and processes it on the traffic lights in time， and has achieved good results in the simulation， further promoting the application of intelligent transportation system.

Dynamic Bandwidth Allocation Algorithm for Passive Optical Networks Based on Utility Function

ZHOU Jiang, CHEN Yang, YU Ling-yun

2023, 0(04): 106-110.

Asbtract ( 83 )

PDF (1240KB) ( 28 )

References | Related Articles | Metrics

In order to solve the problems of long propagation delay and low channel utilization in passive optical networks due to long propagation distance， a dynamic bandwidth allocation algorithm for passive optical networks based on utility function is proposed. In the uplink data transmission direction of passive optical network， the optical line terminal regularly polls each optical distribution network to obtain the request information of uplink bandwidth. In the downlink data transmission direction of passive optical network， the optical line terminal loads the downlink service packets of optical network into frames and transmits data in the form of broadcasting. Combined with this feature， the similarity between bandwidth allocation and microeconomics is used. Based on the transmission power and transmission signal gain， an optimization model of allocation problem based on utility function is established， and the problem model is solved by router matrix and its inverse matrix to obtain the optimal transmission rate of each link and realize the dynamic bandwidth allocation of passive optical network. Experimental results show that the proposed algorithm has high channel utilization and low delay jitter， and can effectively improve the dynamic real-time performance of the network.

A Digital Watermarking Detection Model Based on DWT-SVD and Transfer Learning

CHEN Xiao-wen, SHI Hui

2023, 0(04): 111-117.

Asbtract ( 110 )

PDF (1760KB) ( 22 )

References | Related Articles | Metrics

In recent years， spatial watermarking detection based on deep learning has achieved good results， but the detection result in transform domain is not ideal. To solve this problem， this paper proposes a watermark detection model based on DWT-SVD and transfer learning. The whole model is divided into three parts. In the embedding watermark part， the watermark image is preprocessed first， then the carrier image is processed by three-level wavelet transform and singular value decomposition， and finally the watermark embedding is completed. In the part of transfer learning， the watermarked images and the original images dataset is put into the improved neural network model VGG19-XVGG19， which is used for transfer learning training， features extraction， model parameters optimization， and detection model construction. In the watermark detection part， the model is used to detect and preprocess the image. If a watermark is detected， then DWT-SVD inverse transform is used to extract the watermark. Experimental results show that the proposed watermarking detection model in wavelet domain has short time consumption and high accuracy.

A New Active Defense Model for Complex Network Security

MAO Ming-yang, XU Sheng-chao

2023, 0(04): 118-122.

Asbtract ( 131 )

PDF (3215KB) ( 38 )

References | Related Articles | Metrics

Because most of the existing networks adopt passive defense， which can not effectively resist malicious attacks by hackers in complex environments， an active defense model for complex network security under trusted cloud computing is established. After analyzing the importance of the attacked target and the harm caused by the attack， the attack strategy and defense strategy are classified. By initializing the defense map and judging whether it is 0， and analyzing that the attack and defense game model meets the conditions， the corresponding attack and defense cost is calculated according to the different analysis results. By judging the target and intention of each attack， the corresponding optimal defense strategy is formulated for the defense map. By carrying out simulation experiments and building network topology examples， the results verify that the proposed model can determine the optimal active defense strategy according to the attack target and attack path to ensure the security of network information. It is of great value for information security in a autonomous domain.

Intrusion Detection Method Based on Particle Swarm Optimization Combined with LightGBM

PAN Yu-qing, ZHANG Su-ning, FENG Ren-jun, JING Dong-sheng

2023, 0(04): 123-126.

Asbtract ( 142 )

PDF (1170KB) ( 48 )

References | Related Articles | Metrics

With the development of the Internet， people enjoy the many conveniences it brings， but also face many threats， such as worms and Trojan horses. To defend against these malicious attacks， intrusion detection systems have been created. By detecting anomalies in the current network， intrusion detection systems can effectively detect attacks and take countermeasures. However， the accuracy of traditional machine learning algorithms in intrusion detection models is not high. Based on this， this paper proposes an intrusion detection model based on particle swarm optimization and LightGBM， specifically， an intrusion detection model is constructed by using the LightGBM method and a particle swarm algorithm is used to optimize the parameters of LightGBM. Experiments show that the method proposed in this paper can effectively improve the accuracy of the model， with 98.61% of accuracy， 98.25% of precision， 99.17% of recall rate and 98.70% of F1 score.

Table of Content