Table of contents

The organizing committee of the 5th International Conference on Electrical Engineering, Control and Robotics (EECR 2019) aimed to facilitate interaction among participants for the latest progress and development in the fields of electrical engineering, control and robotics. Through this conference, the committee intends to enhance the sharing of individual experiences and expertise in electrical engineering, control and robotics with particular emphasis on the technical challenges associated with varied applications in these fields.

The conference was held in Xihu Hotel of SCUT, Guangzhou, China, January 12-14, 2019, organized by Sichuan Institute of Electronics with the support of South China University of Technology, University of Electronic Science and Technology of China and Southwest Minzu University. Despite the high quality of most of the submissions, the final proceedings of EECR 2019 includes 62 papers, which were presented at the conference and that were selected after a thorough reviewing process. The authors of these papers come from different countries from Europe, Australia and Asia etc..

The contributions of the technical program committee and the referees are deeply appreciated. Most of all, we would like to express our sincere thanks to the authors for submitting their most recent works and the organizing committee for their enormous efforts to turn this event into a smoothly running meeting.

We sincerely hope that this publication will prove to be an important resource for the scientific community.

Prof. Chun-Yi Su Concordia University, Canada Jan. 18, 2019

All papers published in this volume of IOP Conference Series: Materials Science and Engineering have been peer reviewed through processes administered by the proceedings Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

It is of great significance to ensure the safe and reliable operation of the power grid by effectively evaluating the disaster risk of the power grid lines and arranging the sequence of line guard reasonably. A disaster risk assessment method of transmission lines based on TOPSIS is proposed in this paper where six typical natural disasters of the line are considered, the comprehensive risk values of different lines under different disasters combined with the hazards of line accidents is determined, the risk assessment of multiple transmission lines are realized, and its feasibility is verified by applying it to multiple transmission lines in a certain area of China Southern power grid.

The application of IEC61850 standard and the development of smart terminal unit (STU) technology provide a powerful platform for secondary protection research of distribution network. Aiming at the problems of override trip in the protection of the incoming line and feeders on traditional 10kV switching stations, a scheme combining distributed voltage loss protection and overcurrent protection is proposed. The smart terminal unit collects and processes the operation data, realizes data exchange through GOOSE network communication, and cooperates to realize the protection function of the switch stations. In this paper, a protection logic diagram based on the GOOSE mechanism is constructed, which realizes the protection function of feeder lines fault lockout incoming line breakers and the bus line fault lockout busbar automatic transfer switch.

Our country has shown the significant features of high voltage direct current (HVDC) including multi-infeed and multi-send. Compared to multi-infeed system, there are a lot of weakness in the multi-send system, such as the grid structure is single and weak, which makes the interaction between AC and DC to a greater extent affect the stability of the multi-send hybrid AC/DC transmission system. In order to evaluating the system strength in multi-send hybrid AC/DC transmission system, for the first time, this paper proposed the general formula and simplified formula for multi-send short-circuit ratio (MSSCR), which accounts for the coupling between multi-dc projects. Considering the control characteristics of rectifier in sending-side and combining the maximum power curve method, the criticality of MSSCR is proposed. Multi-send critical short-circuit ratio (MSCSCR) is deduced by the quasi steady-state equation of rectifier, to determine the multi-send system is strength enough or not. The proposal of MSSCR is of considerable significance for effectively evaluating the stability performance of the multi-send AC/DC hybrid transmission system, completing the short-circuit ratio theoretical system, and consolidating the short-circuit ratio index status.

Subsynchronous oscillation is an unstable state in power system, which may damage rotors of turbine generators and break the steady operation of the power system. In China, some accidents of subsynchronous oscillations had happened in the power stations where there are HVDC projects nearby. So that it is necessary to study the Subsynchronous oscillation caused by HVDC. After analysis of the principle of subsynchronous oscillation, the mathematical calculation model of the system containing HVDC is built and analysed in method of complex torque coefficient. And the simulation model is established in PSCAD/EMTDC to prove the validation of the method of complex torque coefficient used for the system containing HVDC, which combines calculation and simulation of the method for the system.

In this paper, the continuous switched boost inverter (CSBI) is used to improve permanent magnet synchronous motor (PMSM) control performance through predictive control, aiming at the drawback of the traditional PMSM drive system. Between the inverted circuit and the input DC power, a switched boost circuit is added to form a switched boost inverter circuit to achieve higher voltage gain and eliminate the dead band. Meanwhile, the drive system has higher reliability. Moreover, the paper adopts predictive current control (PCC) to control PMSM, comparing with the traditional field oriented control (FOC), PCC has faster dynamic performance and reduces current ripple. Finally, the simulation and experimental results verify the feasibility of the based on the continuous switched boost inverter of predictive current control for PMSM.

Reactive power control of hybrid DC asynchronous interconnection system has a vital impact on the safety, reliability and high quality of AC/DC power transmission system. The LCC-HVDC unit and MMC-HVDC unit share the same station control system in the hybrid DC asynchronous interconnected system. Firstly, this paper studies the reactive power switching control of LCC-HVDC system, including five subfunctions and control strategies of the reactive power switching control. Secondly, the reactive power auxiliary control of LCC-HVDC system including low load reactive power regulation function and Gamma-Kick function is studied. And the shortcomings of using reactive power auxiliary control are pointed out. Finally, the reactive power control of MMC-HVDC system is deeply studied. The mathematical model of MMC-HVDC control system is established. Reactive power control of MMC-HVDC can be used to replace reactive power auxiliary control of LCC-HVDC to improve the converter utilization and service life

Transmission distance is a key factor in designing a wireless energy transmission system. The change in coil spacing will have a huge impact on the output characteristics of the system.Therefore, this paper theoretically analyzes the mutual inductance model of two coils from the perspective of circuit theory.Studies have shown that coil spacing, impedance, self-inductance, operating frequency, load resistance and many other factors can affect the output characteristics of the system.So, this thesis uses the control variable method to study the relationship between transmission distance and output characteristics, and proposes the optimal output power point.Finally, an experimental prototype was made to verify the feasibility of the method.

Since the structure of series DC power supply system, if any battery cell breaks down, the group of battery cells can't play its due role. The parallel DC power supply system has been successfully developed and applied in recent years. We want to know whether the parallel DC power system is worth vigorously popularizing and applying in the substation in the future? So we made a few special tests, and this paper will compare the series and parallel DC power supply systems with some data in some ways, such as performance, safety, reliability, maintenance quantity, convenience, economic benefits, etc.

According to the poor stability of the output constant current and the control complexity of the existing Inductive Coupled Wireless Power Transmission System, a seriesparallel compensation network is applied to push-pull converter. And the soft switch is realized by configuring the parameters of compensation network, and the constant-current output of the wireless power is realized without additional complex control system. Firstly, analysis of the push-pull converter to achieve the conditions of soft-switching. Secondly, based on the leakage inductance model of the transformer, combined with the Thevenin equivalent theorem, the parameters of the string-and-compensation are designed. Lastly, the push-pull soft-switching inductive coupled power transfer converter is designed and built.The feasibility of theoretical analysis and parameter design method are verified by the experimental results.

Three-phase unbalance is one of the power quality problems. For the power quality problem caused by the three-phase unbalance phenomenon on the load side in the low-voltage distribution network, combined with the characteristics of the three-phase four-wire system of the distribution network in China, the three-phase unbalance adjustment is studied. The device control strategy is selected, and the three-phase four-bridge arm topology and the three-phase current real-time detection method based on instantaneous reactive power theory are selected. The experimental model is built by MATLAB/Simulink to verify the feasibility of the method.

This paper describes the development of a machine learning model which can be used to screen SiC MOSFETs for paralleling with minimized transient current imbalance. The spread of device parameters is determined to isolate the device parameters which are likely to significantly influence transient current distribution. A linear regression model is then developed and trained using device parameters' data measured from forty devices of the same production lot. The resulting model is an expression for current imbalance as a function of device parameters, with weight of each parameter determined. The performance of the trained model on a set of twenty testing devices is then determined and verified via a double pulse test experiment incorporating paralleled devices. The model is found to perform satisfactorily and can effectively be deployed in screening devices for paralleling.

The voltage transformer is a common nonlinear inductor in the power system. The core nonlinearity caused by saturation, hysteresis and eddy current makes it difficult to describe the dynamic magnetization of the core. At present, there are many different saturation models for nonlinear inductors. Different models are used to simulate ferromagnetic resonance, which will result in different results. In this paper, based on ATP-EMTP electromagnetic transient simulation software, the simulation of different voltage transformer saturation model is carried out, the calculation error is analyzed, and the results of different models are compared.

The existing four-division management method for line loss in smart grids has problems such as the inability to verify line loss contemporaneity and low application flexibility. Considering that the rich data resources provided by data acquisition devices such as measurement meters in the smart grid are not fully utilized, an arbitrary partitioning line loss calculation method that fully utilizes measurement data resources is proposed. The method classifies the grid nodes according to the type of measurement data, obtains complete line power data, and then formulates the boundary equivalence principle of grid partition to realize arbitrary region partition of the power network. On the basis of network partition, the power loss of each subnetwork is calculated by using the power flow of matpower, which greatly increases the flexibility of smart grids line loss management. Taking a power grid in a certain area of Gansu as a case study, the effectiveness of the calculation method is verified.

In this study, it is emphasized that simulation of the avoidance of unnecessary energy cuts in a sample electricity distribution network. The electricity cut-off costs can be decreased by ensuring that relay coordination end only at the source of the fault from the point of failure. Over currents are calculated and the sample network is simulated in the LabVIEW environment. The parameters of delayed overcurrent trip setting (I>), instantaneous overcurrent trip setting (I>> ), delayed ground overcurrent trip setting (Io>), instantaneous ground current trip setting (Io>>) are set according to a specific logic in relay coordination. As a result, you can get an idea about how much profit can be made by doing the relay coordination correctly and the amount of unnecessarily avoided energy. It is aimed with the application of this simulation in the real field that electrically operated materials of the people, who are unnecessarily energy-consuming, have less damage and that the energy-selling company loses less money.

The existing anti explosion measures and the intrinsic safety evaluation methods have become the bottleneck factors that restrict the further improvement of the output power of the native security. Therefore, This article is based on the latest anti explosion standard, the principle of realizing high power output by Power-i technology to suppress the spark discharge energy of switching power supply is analyzed, as well as the composition of the power supply system and its evaluation method. The forward converter is selected as the topology of the high-power local security power supply, according to the technical index of Power-i system, the circuit structure of high power intrinsic safe converter is put forward, and the double protection circuit of overcurrent, overvoltage and short circuit when the output is short circuited and open fault is designed. It has been tested that the output 24V/2A high power intrinsically safe positive converter based on Power-i standard can meet the requirements of explosion protection under type I environment. The experimental results verify the intrinsic safety of the designed high power intrinsic positive shock converter, and reflect the correctness of the theoretical analysis and design method.

This paper analyzes the traction power supply system and traction substation of electrified railway. Due to the use of a large number of power electronic devices, the electric locomotive has strong nonlinear characteristics. Focusing on harmonics, this paper studies the harmonics generated by AC-DC and AC-DC-AC electric locomotives respectively. Based on this, The mathematical model of the power supply system of electrified traction locomotive is established, and the harmonic model is established. The simulation results are in good agreement with the actual conditions.

The efficiency improvement of switching converters has always been a research hotspot. This paper analyzes the loss of LLC resonant converter including magnetic components loss and Switching loss. For easier analysis, a mathematical model of the switching tube is established, and then derive the formula of the loss. The synchronous rectification technology is used to improve the efficiency of the LLC, and finally established a simulation model to verify the correctness of the technology to improve the efficiency of the converter.

Accurately forecast interruptible load can help to alleviate the power supply tension during peak load and make scheduling more flexible. Support vector machine (SVM) method which has been widely used in load forecasting usually selects a period of date close to the forecast day without considering the information characteristics of itself. An interruptible load forecasting method based on clustering algorithm is proposed in this paper. This method puts forward a new idea to select the sample of prediction model which takes full account of the weather and date information of the forecast day and solve the problem that the traditional SVM method cannot properly reflect it. In this paper, the principles of clustering algorithm and support vector machine are introduced firstly. Then K-means clustering algorithm is used to classify the historical data, and the support vector machine forecasting model is constructed by using the categories of the forecast day membership. Finally, the prediction is carried out by combining with the actual data. The results show that the prediction accuracy of this method is more than 95%, and it has higher precision.

The sliding mode observer (SMO) can effectively achieve the permanent magnet synchronous motor (PMSM) to perform position sensorless control, but there are still drawbacks of system chattering and high complexity. In this paper, an improved SMO is established based on hyperbolic tangent function and Kalman filter. The system chattering is effectively suppressed, and the accuracy of position and speed observation error is significantly improved without phase compensation. It is found that the convergence rate of the hyperbolic tangent function has a strong correlation with the system chattering degree and the speed observation error. When a is between 3 and 4, the speed error accuracy of the improved SMO is the optimum situation.

Due to the different manufacturing process and control mode, the commutation torque ripple will occur in the operation of BLDCM(Brushless DC motor). In this paper, the problem of BLDCM torque ripple is analyzed theoretically. In order to suppress the torque ripple during commutation, a method of combining Cuk converter with power inverter circuit is proposed. By changing the bus voltage, the DC bus voltage can be reduced at low speed and increased at high speed, so that the torque ripple of BLDCM can be suppressed at full speed. The simulation results are verified by MATLAB/Simulink, and the simulation results show the effectiveness of the method.

According to the different cooling medium, the electric motor cooling styles can be divided into liquid cooling and air cooling; the former involves water cooling and oil cooling while the latter comprises natural air cooling and forced air cooling. Whereas, the air cooling system is difficult to meet various performance requirements for the modern Electric Vehicle (EV). Therefore, approximately all EV motors are liquid cooled at present. Moreover with the trend of EV motors to Permanent Magnet (PM) and to high speed, for instance, the maximum working rotated speed from 10,000rpm to 20,000 rpm, the packaging of the motor is minimized at same power while efficiency drops, lifetime shorten and even cause demagnetization due to thermal effects without optimum design. Consequently, the improvement of the liquid cooling system is essential to enhance the efficiency of the EV motor in the future for higher power density. In this perspective, the innovative cooling topologies of electric machines based on the development results of the parts such as lubricating oil and motive seal of electric vehicles are presented in this paper. It solves some problems brought about by predecessors in view of the high efficiency and reliability of the motor with higher energy density.

In order to improve the control precision of permanent magnet synchronous motor servo drive, a Proportional-Integral-Differential (PID) control method based on particle swarm optimization (PSO) algorithm is proposed. Firstly, the mathematical model of the permanent magnet synchronous motor was established. Then the PSO algorithm was used to adjust the parameters of the PID controller. Finally, the simulation experiment was carried out under the MATLAB/Simulink environment to test its performance. Compared with the traditional manual parameters tuning, the PID controller based on PSO algorithm had better performance and better stability.

The design of an fuzzy force based on grasped objects stiffness method for tendon-driven underactuated prosthetic hand system are addressed in this study. First, a mathematical model of tendon-driven underactuated prosthetic hand is developed for force control. Next, fuzzy force based on grasped objects stiffness was discussed to improve the control performance. Finally, the proposed control strategy for underactuated prosthetic hand is simulated to six different stiffness objects. The results shown that the controller is suit for prosthetic hand grasp objects with different stiffness. The simulation results verify the effectiveness of the force control method based on grasped objects stiffness.

In this paper, the principle of model reference adaptive method (MRAS) is analysed and studied. Aiming at the shortcomings of installing speed sensor to collect the speed signal and feedback to the control system, a method for speed estimation and parameter identification of asynchronous motor without speed sensor is presented, And taking the model reference adaptive method (MRAS) as the main line, the Popov super stability theory is used to derive and analyse the adaptive law. Finally, the proposed method is verified by MATLAB software and hardware circuit experiments.

The five-axis welding robot based on leadshine control system can accomplish various welding tasks in industry, and can accomplish arc welding and linear welding and other welding types. When welding, the machine has small chattering and high welding accuracy. In this paper, the D-H linkage coordinate system and kinematics forward and inverse solutions of the robot are analyzed, and the reachable space of the robot is calculated by Monte Carlo method. The hardware system of the robot is also summarized in more detail in this paper. Using SimMechanics tools to model the robot. Then using MATLAB function to validate the model, recording the difference between theoretical data and simulation data, and judging the correctness of the theory according to the difference.

Unmanned Aerial Vehicles (UAV) are becoming popular to the innovating world. This paper presents a novel experiment design to determine the performance of UAV control system. The hardware platform developed is composed of a quadcopter with a multipurpose autopilot system. While the experimental design considers both quantitative and qualitative aspects of the controller. The qualitative components dealt with the hardware compatibility, acceptability and license. On the other hand, the quantitative component measured the accuracy of waypoints, altitude hold and loiter performance of the UAV. The results shows that the new experiment design platform was able to compare two popular autopilot software in terms of performance and acceptability.

In this article, development of an autonomous underwater vehicle (AUV) path planner based on unstable mode is considered. The path planner is based on the bionic approach and does not require pre-mapping, it reduces requirements for the sensor subsystem of AUV. In the path planner, we use a method for obstacle avoidance in which an unstable mode is realized in a separate dynamic link. The output of the dynamic link corrects the desired angles of the AUV orientation. It allows setting the unstable mode only at the planning level, and at the regulatory level, a control system works in a stable mode. In addition, this approach allows planning the direction of obstacle avoidance. Detailed description of the proposed path planner and results of the study in Matlab are provided.

We propose a distributed approach to multi-aircraft real-time conflict resolution based on satisficing game theory which considers reducing flight collisions and improving economic efficiency. In our conflict game model, aircraft as exclusively self-interested agent are assigned different statuses by a priority ranking mechanism proposed. According to the state of one's own priority, they take others' preferences into consideration to choose the most satisficing conflict resolution strategy from policy set including speed and heading modification strategies. The simulation results on the random flight scenario of two concentric circles show that our method which adopts speed and heading adjustment to avoid conflicts is promising in terms of safety and system efficiency.

As container terminals aim for more efficient operation and higher productivity, automation is making major breakthroughs all over the world. Container terminals are increasingly adopting automated solutions to meet the challenge of larger ships, taller cranes and bigger call sizes. Having the right design criteria and taking the correct decisions early in the project is crucial in order to meet the expected cost and performance targets. Therefore, designing an automated container terminal is no longer just a matter of acquiring the right port equipment and integrating the equipment with basic terminal operating systems (TOSs), it is becoming more of designing and implementing a complete smart automated system. Smart automation would enable Automated Guided Vehicles (AGVs) used in container terminals to automatically adapt to changing circumstances, optimizing performance and output. Most of existing research work considered pre-defined AGVs paths using different techniques without considering free path AGVs solutions that can potentially increase transportation efficiency This research work uses the concept of Internet of Things (IoT) to increase transportation efficiency of the Automated Guided Vehicles (AGVs) and reduce vessel discharging time in an automated container terminal. In addition, the research paper proposes an algorithm to increase efficiency in the automated container terminal and achieve the research objective. Finally, simulation results were used to prove the positive effects of using the proposed algorithm in decreasing vessel discharging time compared to existing benchmarks and optimum solutions.

In the development of railway transportation, safety has always been the most concerned issue. Balise is the core content of high-speed railway train control system, it is one of the most basic unit of information storage and transmission, control and status information necessary for the train. It is affected by many factors in the process of operation, so it is unavoidable to have problems of malfunction. At present, staff are required to operate the transponder to repair or update message information. In order to meet the needs of daily maintenance, the scheme of reading and writing operation of transponder by dynamic train detection is proposed and its feasibility is analyzed. This paper designs the software part of the mobile detection system by analyzing the working status information of different versions of transponders and tests it. The test results show that the dynamic reading and writing operation scheme is feasible.

As the planning and construction of electric vehicle charging pile plays a decisive role in the promotion of electric vehicles, this article puts forward a planning method considering the cooperation of different types of charging piles and analyzes the Residential and social charging piles. On the basis of determined number of charging piles in residential area, the planning of social charging piles is analyzed from the demand of charging considering the unbalance of electric vehicles' promotion, the location of charging piles considering the minimum charge distance and the capacity of charging piles considering the situation that EVs don't charge every day. And then the processes of planning are determined. Finally, the planning method is applied to a Xincheng planning in ZhuoZhou which could illustrate the feasibility of this method. Compared with other method, the method of this article pays more attention to the actual situation from the load forecast to the overall planning in the area, thus the result is more realistic and in line with the planning requirements.

In order to evaluate the degradation state of the mechanical equipment and master the information of the remaining useful life (RUL) of the bearing accurately, this paper presents a method for predicting the remaining useful life of bearings based on mutual information (MI) and support vector regression (SVR) model. The proposed method includes two steps of online and offline, the offline step is used to build a degradation model of the bearing by learning, the online step uses the degradation model to predict the remaining useful life. By analyzing the experimental data of bearing full lifetime degradation, the results show that the method can effectively simulate the bearing degradation process and predict the remaining useful life of the bearing.

In order to solve the problems such as small workspace, complicated mechanical structure and motion interferences of traditional exoskeleton arm, a modified exoskeleton mechanism is proposed by optimizing the angles between the shoulder joint axes. In this paper, the model of traditional exoskeleton arm and the modified one are both built. And the kinematic theory of the two models is analysed and discussed. Furthermore, the Monte Carlo method is used to simulate the workspace of the mechanism with the help of the Matlab software. The simulation results show that the modified exoskeleton arm has larger workspace than the traditional one. So it can better meet the needs of more application scenarios.

With the large-scale application of electric vehicles, the construction of electric vehicle charging stations has become a prerequisite for urgent needs. In this paper, TLBO optimization algorithm is proposed. The optimal model of the location and the constant volume of the charging station is established based on the weighted distance and minimum distance from the user to the charging station. Based on the layout of the district network in the planning area, the construction plan of the electric vehicle charging station in the planning area is studied. Through a case study of a planning area, the results show that the TLBO optimization algorithm is applied to the location and constant volume of electric vehicle charging stations, with accurate calculation, rapid convergence, and good global optimization performance.

In order to predict the distribution of gas flow, we need to get the temperature of each point in blast furnace throat in advance. In this paper, firstly, two intelligent modeling methods are used to establish a multiple-input multiple-output prediction model, one is extreme learning machine (ELM) algorithm and the other is online sequential extreme learning machine (OS-ELM) algorithm. And the model is a single-step prediction model of temperature in blast furnace, single-step prediction means the prediction of temperature in the next moment. We use autocorrelation analysis to determine input vector and output vector of the model. The result of autocorrelation analysis indicates that the method of temperature sequence prediction has a higher prediction accuracy and better prediction stability than the method of single point prediction. Next, based on real industrial data, we make a comparison between the multiple - input multiple-output model and least squares support vector machines (LS-SVM) model used in common. The experiment results show that OS-ELM model has a better forecast effect than the ELM model and LS-SVM model.

With the development of switching power supply towards miniaturization and high efficiency, digital and soft switching technology has become the research focus of switching power supply. Phase-shifted full-bridge DC/DC converter is commonly used in medium and high power applications due to its advantages of low switching loss, high efficiency and low output current ripple. Based on the existing switching power supply and digital control technology, a digital control phase-shifting full-bridge switching power supply is designed. Its output voltage ranges from 0 to 12 V.The output current ranges from 0 to 50 A. The voltage and current level can be adjusted according to the user's power demand, and the output mode of steady voltage or steady current can be set. The parallel operation of multiple power supply is realized by RS485 communication protocol.

Camera tracking is an important application in the computer vision. With the progress of the tracking-by-detection algorithm, industrial cameras can track targets autonomously. However, during the long-time tracking, it's prone to miss target owing to heavy object occlusion, environment changing and objects appearance variations. Our camera tracking system is based on the kernelized correlation filter to track object, and we add a self-verification module to judge if the current tracking results are reliable. This can be useful in solving target missing when object meets occlusion or variations, and avoid model drift during long-time detection. Last but not the least, we keep the targets in the corner of the image by controlling our camera, avoiding object's moving out of view. The kernel correlation filter bounded with self-verification and re-detection modules boost the performance. Extensive experiments on the OTB-2013 benchmark show that it performs better than state-of-the-art methods.

To solve the issues including the inflexibility, connector wear, and electric shock risk in contact charging system for inspection robots, the wireless charging system for inspection robots is introduced in this paper. The circuit model of this paper is developed based on the mutual inductance theory. The expressions of the receiving power and the efficiency are derived. The coordinate calculation method for mutual inductance between non-coaxial rectangular spiral coils according to the movement characteristics of inspection robot is presented. According to the calculation method, the characteristics of the receiving power and the efficiency varying with the position change of the receiving coil are investigated. The location of the positioning device for starting the wireless charging system can be selected based on the system characteristics. The theoretical analysis is verified by experiments.

The forward-flyback converter is used in power factor correction to improve the performance of conventional forward PFC and flyback PFC converters. By analyzing the working principle of a single-transistor single-transformer forward-backward combined PFC converter in one power frequency cycle, it is pointed out that there are two working modes — flyback mode and forward flyback mode; through its PFC mechanism research. The relationship between the duty cycle of each operating mode when achieving unity power factor is obtained, and the duty cycle when operating in flyback mode is the maximum duty cycle within a power frequency cycle. Finally, the open-loop control model of the forward-flyback PFC converter is established by using MATLAB. The duty cycle is changed according to the specified rules and the PFC is implemented, which verifies the correctness of the theoretical analysis.

In order to solve the problem of poor convergence and low precision of the current Multi-sensor Detection Network algorithm, a multi-sensor cross cueing technology based on Grey Wolf algorithm is proposed. By endowing the grey wolf with autonomous decision-making ability, the algorithm can jump out of the local optimal solution as soon as possible, and the convergence degree of the algorithm is improved. The simulation results show that the algorithm can find the optimal solution in a short time and maintain good stability, and the optimization ability is further strengthened. By comparing with other algorithms, it is more effective to prove that the algorithm has better solution quality.

In the case of terminal fault of long line, the failure protection of traditional circuit breaker has the problem of refusing to operate due to the insufficient sensitivity of voltage locking elements. A new circuit breaker failure protection scheme based on station area information sharing is proposed in this paper. The new scheme will replace the protection export signal with the actual protection criterion. In addition, it sets up two current transformers to detect whether the breaker is cut off after the circuit breaker is cut off, and checks the redundant current of the bus, which can improve the accuracy of the current detection. Besides, it uses reclosing checkout voltage device signal to supplement the original compound blocking signal. Theoretical analysis and simulation results show that the scheme can effectively cover the full length of the transmission line and has high reliability.

Geomagnetically induced current (GIC) caused by geomagnetic storms flows between the power grid and the earth would threaten the stable operation of the power system. Capacitor blocking device connected in series at transformer neutral point can directly block the DC path. However, due to the complexity of the DC path in the power system, before the installation of the blocking device, if global considerations are not taken into account, the adverse effects of GIC will not be reduced. Therefore, to study the position optimization of the capacitor blocking device is of great value. This optimization of placement of capacitor blocking device is proposed in this paper. Applying the idea of genetic algorithm, the installation quantity and installation position of the blocking device were calculated and compared with the traditional GIC non-optimal method, which verified the superiority of the genetic algorithm optimization solution.

In order to promote vehicle active safety and to decrease collision accident, In this paper, an algorithm for detection of vehicle distance in front based on single and double camera switching is presented. This algorithm has done the corresponding difference processing to the day and night two different situations. By switching between single and double cameras, the image of the car bottom shadow, the characteristics of the vertical direction of the vehicle, the characteristics of the car taillight and other features was taken, and the position of the vehicle in front was locked. Finally, the feature information of the vehicle is converted into the distance information. Experiments show that, this system designed in the paper, achieved the anticipated function, can accurately detect multi-lane and short distance vehicle distance, and the efficiency of the system also achieved the real-time requirement.

Recommender systems have become an essential part in our life. Most social Websites use recommender systems to enhance user experience. In online shopping Websites such as Amazon, Clothing is one of the most popular domains, therefore a recommender is of great significance. Previous recommender systems were often focused on retrieving items based on user preference, i.e. similarity to the previous items purchased by the user. However, in Clothing domain, the matching relationships between candidate items and the previously purchased is also important for recommendation. For example, a user may want to buy a new jean rather than suit pants if he/she has just purchased a shirt. This kind of matching relationships also frequently occurs in other life contexts. In this paper, we aim to recommend new clothes that can better match the clothes purchased by a user. This new recommendation strategy would work better in the Clothing domain and complement the current recommendation literature. Experiment results show that our method can lead to better recommendation performance in the Clothing domain.

The traditional CNN extracted image features is insufficient, so the classification accuracy is not ideal. It also brought overmuch model parameters and calculation. Based on these problems, this paper proposes a dimension reduction residual module. First, reducing the dimension of the output feature map. Then, feature extraction used two different sets of convolution kernels. It can get more sufficient and different characteristic information. Two sets constitute the cascaded layer. Finally, the cascaded layer act as the input of the next layer. This module can reduce parameters. Meanwhile, it increases the depth of the network and enriches the diversity of feature acquisition. A new convolution neural network is build through this module. The performance of new network and other recognition algorithms is compared on GTSRB and 101_food datasets. The new network model is reduced to about 6.2MB, and the classification accuracy can reach 98.2% on GTSRB, 72.3% on 101_food. The experimental results show that this module can effectively improve network performance and control model size.

With the explosive development of mobile Internet, short text has been applied extensively. The difference between classifying short text and long documents is that short text is of shortness and sparsity. Thus, it is challenging to deal with short text classification owing to its less semantic information. In this paper, we propose a novel topic-based convolutional neural network (TB-CNN) based on Latent Dirichlet Allocation (LDA) model and convolutional neural network. Comparing to traditional CNN methods, TB-CNN generates topic words with LDA model to reduce the sparseness and combines the embedding vectors of topic words and input words to extend feature space of short text. The validation results on IMDB movie review dataset show the improvement and effectiveness of TB-CNN.

As one of the most important component of artificial intelligence, machine learning is getting more and more attention. AdaBoost, a classic machine learning algorithm, is widely used. However, when faced with imbalanced data classification, AdaBoost's recognition rate of minority samples is low due to ignoring class imbalance. In many cases, minority samples are of high value. For this shortage, combining the theory of margin and cost-sensitive idea, a new Boosting algorithm called CMBoost is proposed based on cost-sensitive margin statistical characteristics, which is firstly through optimizing margin statistical characteristics to improve formal algorithm and then extended by cost-sensitive. Experimental results on the UCI dataset show that the CMBoost algorithm is superior to AdaBoost for imbalanced data classification problem.

Influence maximization is the problem of finding a small subset of nodes (seed nodes) in a social network that could maximize the spread the influence. The algorithm problem of computing influence estimation and influence maximization have been extensively studied for decades. In this paper, we studied the reverse influence sampling method proposed by Byogs et al. in 2014, and some other improved algorithms. We proposed a new algorithm TESA (Time-Space Efficient Algorithm). On large social networks, TESA saves more space than existing algorithms and runs faster. We used data sets tested in the literature to evaluate TESA experimentally and showed that it was superior to the most advanced solutions in terms of running time and memory footprint.

In recent years, green buildings have developed rapidly in China, but there are still many problems in its operation and maintenance management, it is difficult to meet the requirements of green buildings' operation. In view of the complexity of space, function and management mode, the problem of the operation and maintenance management of green buildings is more prominent. An intelligent operation and maintenance management cloud platform for green buildings of group-company was designed and implemented based on IOT technology in this paper. The platform can significantly reduce the difficulty of the operation and maintenance management of green building groups and reduce the number of operation and maintenance personnel and energy consumption, and reduce operating costs.

Background: In the field of diagnostic CVD, the predecessors used a large amount of data with no missing two-category data, and obtained good results. However, in the process of electronic input of historical data, a large number of data attribute values are missing, and there are multiple levels of disease risk. Goal: On the data set of imbalance and a large number of missing values, this paper focuses on the five levels of cardiovascular disease. Methods: A new prediction model of Adaboost+RF is constructed by using the information gain ratio to analyze the feature contribution degree of the data set. The performance of this model is evaluated with Precision, Recall, F-measure and ROC Area values. Results: The results show that the four key indicators of the Adaboost+RF model on five-categories unbalanced datasets in Precision, Recall, F1 and AUC values, which are 40.9%, 49.3%, 41.4% and 71.6%. Conclusion: The experiment results demonstrate that the four key indicators of the Adaboost+RF model on five-category unbalanced missing datasets are better than other machine learning algorithms.

Nearest neighbor search is playing a critical role in machine word translation, due to its ability to obtain the lingual labels of source word embeddings by searching k Nearest Neighbor ( k NN) target embeddings from a shared bilingual semantic space. However, aligning two language distributions into a shared space usually requires amounts of target label, and k NN retrieval causes hubness problem in high-dimensions feature space. Although most the best-k retrievals get rid of hubs in the list of translation candidates to mitigate the hubness problem, it is flawed to eliminate hubs. Because hub also has a correct source word query corresponding to it and should not be crudely excluded. In this paper, we introduce an unsupervised machine word translation model based on Generative Adversarial Nets (GANs) with Bilingual Similarity retrieval, namely, Unsupervised-BSMWT. Our model addresses three main challenges: (1) reduce the dependence of parallel data with GANs in a fully unsupervised way. (2) Significantly decrease the training time of adversarial game. (3) Propose a novel Bilingual Similarity retrieval for mitigating hubness pollution regardless of whether it is a hub. Our model efficiently performs competitive results in 74min exceeding previous GANs-based models.

As the mobile internet playing an increasingly important role in people's daily life, the copyright issue of Android applications attaches more and more attentions. Because Android applications are vulnerable to reverse engineering, there are many criminals illegally modify and re-sign the official android for their own interests, which has seriously damaged the user's rights and disturbed the order of the mobile internet. In this paper, a novel detects method of Android applications similarity based on the authors' programming style is described. In the proposed method, we firstly extract a series of basic features that can represent the programming style of author. Considering that the features selected manually are highly subjective, we tried ICA, PCA and LDA to optimize the features. Ultimately, an Extreme Learning Machine (ELM) model is trained with sequences of feature vectors. The results of experiments demonstrate that our proposed method has better performance than string comparison and grammar tree analysis when fighting against variable substitution, insert independent statement and data stream confusion.

The segmentation of retinal vessels is greatly significant for doctors to diagnose the fundus diseases. However, existing methods have various problems in the segmentation of the retinal vessels, such as insufficient segmentation of retinal vessels, weak anti-noise interference ability. Aiming to the shortcomings of existed methods, this paper proposes an improved model based on the U-Net networks, which contains densely-connected convolutional network and a novel attention gate (AG) model, referred as Densely-Attention-U-Net (DA-U-Net), to automatically segment the retinal blood vessels. The method can alleviate the vanishing-gradient problem, strengthen feature propagation, substantially reduce the number of parameters, and automatically learn to focus on target structures without additional supervision. By verifying the method on the DRIVE datasets, the segmentation accuracy rate is 96.09%, higher than that of U-Net and R2U-Net.

This paper proposes a CNN-based (Convolutional Neural Network based) network to detect altered face picture, which can cover the most common face swap methods. The network uses an autoencoder which is pre-trained on the original images to reconstruct the input images. The reconstructed one and the input image are then processed by the SRM filter which can extract the noise distribution of images. We then feed the minus result of two processed results into a CNN architecture to predict whether the input image is original or tampered. The model was trained and evaluated in FaceForensics dataset and state-of-art face swap method. Experimental results demonstrate the effectiveness of our network.

A method and software tool for measuring and reconstructing small plant in real time was presented. The key of this method lies in the feature points measuring with 3D digitizer and organ templates-based details 3D reconstructing. Firstly, several feature points were selected and measured by using a digitizer from the targeted plant. Then the morphological parameters of this plant could be calculated from these feature points, with its skeleton model was reconstructed in the process of measuring. For a details 3D reconstruction of the plant structure, organ's 3D template was used, which was created by using 3D point cloud data measured from real plant organs, or created with interactive designing. This method has been tested in many plants, including crops and small fruit trees. The experiment results indicated that the proposed method is suitable for real-time measurement and multi-scale structural modeling of small plants.

Accurate and rapid detection in remote sensing images is of great significance in military navigation, environmental monitoring, and civil applications. Due to the small object detection problem of remote sensing images, higher requirements and challenges are put forward for object detection technology. In recent years, the one-stage object detector and the two-stage object detector based on convolutional neural network have made great achievements in the field of image classification and detection. The one-stage object detector is generally superior to the two-stage object detector in detection speed, but the detection accuracy is inferior to the two-stage target detector. In this paper, YOLOv3 is used as a one- stage target detector for small aircraft detection of remote sensing images. We chose appropriate anchors to cover the size distribution in our experimental data by using dimension clusters. The experimental results show that YOLOv3 not only exceeds the conventional one-stage object detector in speed, but also well match the accuracy of the two-stage object detector. The YOLOv3 achieved excellent detection accuracy and low processing time at the same time for small aircraft detection.

The traditional Structural SIMilarity (SSIM) index has higher computation complexity compared with Peak-Signal-to-Noise-Ratio (PSNR). In order to solve this problem, we propose a two-stage optimization algorithm to reduce the computation complexity of SSIM. First, floating-point operations in the calculation process of SSIM are converted to fixed-point ones. Then, the repeated calculation rule of the iteration operation in SSIM is analyzed and the repeated operations are subsequently removed via modifying SSIM calculation method. Experiment results show that, in comparison with the original SSIM algorithm, the proposed algorithm obtains 24.12% of computation time reduction with negligible accuracy decrease.

In recent years, deep learning has made great progress in computer vision, and shown a good application prospect in reading medical images. TAO (thyroid-associated ophthalmopathy) is one of the most common orbital diseases in adults is autoimmune disease, the exact pathogenesis is not clear. In hospitals, CT images are usually used for disease analysis and the distortion of ocular rectus muscles is one of the main causes of TAO, thus the separation of the ocular rectus muscle is of great practical significance. However, the artificial segmentation of the ocular rectus muscle is a time-consuming task and relies heavily on the experience of the operato. In this paper, we improved the traditional U-net with GoogLeNet inception module and apply it on the segmentation of ocular rectus muscles. The experimental results show that the segmentation results of our method are more accurate and efficient than the traditional algorithm, and it is helpful for doctors to diagnose patients' diseases.

The traditional incremental learning tracking algorithm tends to drift or lose track when the pedestrian is occluded and the background is similar to the pedestrian target. This paper proposes a pedestrian tracking algorithm based on improved incremental learning. According to the difference between the central position of the current observation model and that of the previous one, the reliability of the observation model can be judged, and then the forgetting factor can be adjusted, so that the observation model can adapt to the changes of pedestrian targets. The experimental results show that the improved algorithm can track the pedestrian target accurately and effectively in the occlusion scene.

In this paper, an approach based on combining surface electromyography(sEMG) and three-axis acceleration(ACC) signal was proposed to recognize 5 different kinds basis daily gait patterns, including walking on the ground, going up stairs, going down stairs, going up slope and going down slope. Firstly, the gait related sEMG signal and three-axis ACC signal were collected from the lower limbs of subjects. Secondly, the de-noising of sEMG signal was finished and the segmentation of the fusion signal was done. Thirdly, the features of fusion signal were extracted. Finally, a classifier based on 2-stream hidden Markov model (HMM) was built to recognize 5 kinds of basis daily gait patterns. The experiment obtained an average recognition accuracy of 94.32%, which is 4.15% higher than the accuracy by adopting sEMG signal only (Average 90.17%), and 9.60% higher than the accuracy by adopting ACC signal only (Average 84.72%). The result demonstrated that it can improve the recognition accuracy of gait patterns effectively to combine sEMG signal and three-axis ACC signal.

Segmentation of Femur bone from X-ray images is an indispensable step in computer aided analysis of medical images and orthopaedic examinations. It is more complex than segmentation from CT and MR images, due to some associated less dense tissues that are hard to distinguish from the femur bone in X-ray images. This paper presents an improved method based on U-Net to automatically extract the femurs from hip X-ray images. This method changes the structure of the U-Net network, which can effectively map the non-linear relationship between hip image and femur image, and accurately segment femur image. The paper also added the absolute deviation loss function to improve the segmentation effect. Experimental results show that this method is accurate, robust, and achieves an average dice similarity coefficient of 0.966. The segmentation results are satisfactory.

Although the state-of-the-art object detection methods, which depend on region proposal algorithms to hypothesize object locations, have achieved high detection accuracy, it still struggles in small-size object detection. In this paper, we present a novel method with a multi-scale and multi-tasking region proposal method to effectively detect small object. In the proposed method, multi-scale features and high-level features are employed to locate object position and identify object category, respectively. The main contributions of the proposed approach are two-fold: (1) A simpler way is used to improve the accuracy performance of small object detection, instead of the complex image pyramids and the complex combination framework, and make the object detection task more flexible. (2) Based on multi-scale and multi-tasking approaches, object location information in low layers and object semantic information in deep layers are made fully advantage respectively. The experimental results on the PASCAL VOC dataset show that the proposed method achieves the state-of-the art object detection accuracy.