Abstract：Deterministic technology is an important research content to meet the deterministic forwarding requirements of 6G emerging services. As the kernel system of Industry 4.0, the deterministic guarantee of Cyber-Physical System(CPS) is of vital importance for the efficiency and stability of production and the safety and reliability of service. Focusing on the deterministic requirements of CPS represented by industrial Internet, a CPS system architecture based on deterministic network technology is proposed, including deterministic communication network and deterministic system components. On this basis, the proposed system architecture is detailed designed for three typical use cases: low delay cloud control scenario, high reliability mobile control scenario and large bandwidth closed-loop control scenario, and the determinacy of delay, jitter, reliability and bandwidth is analyzed.

Abstract：With the development of mobile communication technology, the 6th Generation mobile networks(6G) network will ease the pressure on the bearer network through distributed network sinking, and optimize resource allocation through distributed coordination. However, how to ensure distributed trustworthiness and help the multi-party collaborative system is an upcoming security challenge. As a distributed ledger technology, blockchain can be applied to cross-domain authentication and cross-domain resource coordination. Blockchain can ensure the transparency of transactions and the credibility of data, and establish a trust relationship between multiple participants, which become the key technology to solve distributed trust.

Abstract：With the development of delay-sensitive services such as telemedicine and smart factories, the demand for mobile deterministic networks continues to increase. The 3GPP R16 version proposes the 5G TSC network architecture of 5G coordinated TSN, which realizes the deterministic service of the mobile network. However, the 5G TSC network architecture still faces many technical problems: not supporting joint service scheduling, not supporting long-distance WAN transmission, and not supporting application collaboration. Therefore, this paper sorts out the promotion process of 3GPP supporting mobile deterministic networks, clarifies the reasons for the existence of problems, and provides solutions based on artificial intelligence algorithms for 6G networks.

Abstract：The demand of industrial manufacturing and engineering development for simulation analysis and calculation is increasing, the scale of simulation analysis model is larger, the accuracy of the calculation is higher, there are more and more multidisciplinary simulation, all above needs more high performance computing capability. A number of independent high performance computing cluster is not enough conducive to resource sharing and efficiency. This paper presents a design method of high performance computing platform based on multi cluster.

Abstract：In the field of electricity safety, there are a lots of data and knowledge has not been excavated and utilized, constructing a knowledge graph in the electricity safety field can not only integrate power knowledge, but also greatly improve the efficiency of the power industry. Named entity recognition(NER) is the basis for constructing knowledge graph, this paper studies the named entity recognition based on dictionaries and rules, through three methods: the domain entity dictionary, the word-building feature character rule matching and the part-of-speech combination feature rule matching，to accurately extract electricity safety related entities from non-structured text, providing high-quality and high-precision entities for the construction of knowledge graph in the field of electricity safety. In order to optimize the recognition process and improve the response speed, the general part-of-speech tagging task is sent to the edge node for processing, and the central server processes the rule template matching task. Experimental results show that using the three methods comprehensively to recognition the domain entity of small-scale electricity safety text, the F1 score can reach more than 85%.

Abstract：With the rapid development of power transmission and distribution grid, power transmission and distribution services have put forward higher requirements for differentiated communication metrics such as delay, energy efficiency, and reliability. Based on the characteristics of high speed and large connectivity of 5G and flexible and wide coverage of satellite, combined with the advantages of efficient data processing of edge computing, a 5G satellite integrated networking architecture for power transmission and distribution scenarios is constructed. Then，with the optimization goal of maximizing the weighted difference between the average energy efficiency and the total delay of data offloading in the transmission and distribution services, an optimization problem of data offloading in 5G satellite integrated networking is constructed. A differentiated service requirement-aware learning-based data offloading algorithm(DSRL-DO) for 5G satellite integrated networking is proposed to solve the problem. Finally, the simulation results show that the proposed algorithm can reduce the data offloading delay while ensuring high energy efficiency, and effectively meet the differentiated requirements of power transmission and distribution services.

Abstract：In recent years, China′s independent software and hardware ecosystem has developed rapidly. However, the independent technology system still has the pain points of resource scarcity, difficulty in obtaining technical services and high cost. To this end, Pengcheng Laboratory built a service-oriented independent hardware and software ecology based on the concept of open source and openness, built a developer cloud platform to provide developers with integrated independent software and hardware technology stacks and services for free. This paper introduces the construction achievements and services of Pengcheng-ecology in the past three years. Based on the usage data of the developer cloud platform, we calculated the comprehensive economic benefits of the platform. By analyzing the behavioral data of the platform users, we find three development trends of independent industry, which provides a new perspective, new ideas and new methods for the development of independent software and hardware ecology.

Abstract：Energy Internet is the key to realize the reform of energy structure and improve the safety and green level of energy in China. Therefore, the security of the energy Internet is related to the stable and safe operation of the energy Internet. Therefore, we should study the energy Internet security situation awareness system, build a comprehensive monitoring platform for the energy Internet security situation, and realize the monitoring and analysis of the energy Internet security situation and the security risk monitoring of the internal and external networks. Build a global energy Internet security reinforcement and protection system, and form a series of security norms, so as to achieve a comprehensive awareness of the energy Internet security risk situation, efficient early warning of security threats, improve the controllable level of energy Internet security, ensure the energy Internet Security in all directions and from multiple angles, and ensure the vital interests of users.

Abstract：Lip reading is crucial in the silent environment or environments with serious noise interference, or for people with hearing impairment. For word-level Chinese lip reading problem, SinoLipReadingNet model is proposed, the front end of which with Conv3D and ResNet34 is used to extract temporal-spatial features, and the back end of which with Conv1D and Bi-LSTM are used for classification and prediction respectively. Also, self-attention and CTCLoss are added to improve the back end with Bi-LSTM. Finally,the SinoLipReadingNet model is tested on XWBank lipreading dataset and results show that the prediction accuracy is significantly better than that of D3D model, the prediction accuracy and avrage CER of multi-model fusion reaches 77.64% and 21.68% respectively.

Abstract：In order to solve the problem that common target detection algorithms are difficult to apply effectively in classroom scenarios, a student behavior detection algorithm combining lightweight and trapezoidal structure is proposed. The algorithm is based on YOLOv4 architecture, according to the characteristics of target classification and distribution space, a new “trapezoidal” feature fusion structure is proposed, and combined with the MobileNetv2 idea, the model parameters are optimized to obtain a trapezoidal-MobileDarknet19 feature extraction network, which not only reduces the computational load of the network, but also improves the work efficiency. At the same time, it strengthens the information transmission of target features and improves the learning ability of the model. In the scale detection stage, a five-layer DenseNet network is introduced to enhance the network′s detection ability for small targets. The experimental results show that the proposed YOLOv4-ST algorithm is better than the original one. The mAP of YOLOv4 algorithm is improved by 5.5%. Compared with other mainstream algorithms, it has better practicability in the task of student classroom behavior detection.

Abstract：In order to meet the market demand of low-cost, low-power consumption and high linearity of the Internet of Things, a 2.4 GHz power amplifier(PA) with high linearity is proposed. The power amplifier has a two-stage structure. In order to improve the gain while maintaining low static power consumption, the driver stage of the PA adopts a current multiplexing two-stage common source amplifier structure, uses a two-stage distortion cancellation method to reduce transconductance nonlinearity, and adopts diode linearization bias to compensate gain compression phenomenon caused by parasitic capacitance nonlinearity. The PA uses a 0.18 μm CMOS process. Simulation results show that at 2.4 GHz operating frequency, the PA has a small signal gain of 30 dB, an output 1 dB compression point of 21.7 dBm, a static power consumption of 53 mW, and a power-added efficiency peak of 31%.

Abstract：To decrease the costs of the implementation of compression by software in a storage system, the lossless compression circuit based on LZ4 algorithm is proposed in this paper. The sliding dictionary with circulating queue structure is used, together with the Hash Table module to design the lossless compression circuit based on LZ4. Finally, the design is tested and verified with Calgary corpus and Canterbury corpus on Xilinx KC705 FPGA platform. The results are compared with LZ4 software method and it can be concluded that the compression efficiency of LZ4 circuit method greatly faster than the software method obviously, while the compress ratios keep the same.

Abstract：Phase-locked loop(PLL) based on dual second-order generalized integrator(DSOGI-PLL) is widely used in inverter synchronous grid-connection. However, when the power grid voltage is unbalanced or harmonic components and DC offset exists, namely, when the grid voltage condition is complex, DSOGI-PLL has a large deviation from the frequency. Therefore, an improved structure is proposed. The DC component suppression module is added to filter the DC voltage, and the low-pass filter is added to improve the ability of the PLL to suppress harmonic voltage, and accurately extract the grid voltage synchronization signal. The simulation results in MATLAB/Simulink verify that the PLL can filter out the harmonic component and DC component, and the effect is better.

Abstract：Bipolar plate is one of the two important components of Proton Exchange Membrane Fuel Cell (PEMFC). It has a great impact on the overall weight, volume and power of fuel cell, and plays a key role in the overall production cost and production cycle of fuel cell. The graphite composite plate produced by molding can form the whole plate at one time, save the secondary processing time and greatly improve the production efficiency. The design of graphite composite plate molding automatic production line includes six parts: feeding section, molding section, drying section, post-treatment section, detection section and packaging section, which can meet the industrialized and mass production of plates and the demand for plates in the fuel cell industry.

Abstract：Measure and control equipment in our country is a kind of necessary strategy to ensure the smooth progress of Chinese space mission. In recent years, technicians have used software simulation provided by this equipment to do daily training. However, this kind of simulation cannot simulate all the equipment functions. Furthermore, it is difficult to update the simulation program for a new scene. By combining a drone and an electric payload, this paper introduces an algorithm that simulates the movement of a target. The simulation results indicate that the proposed algorithm can calculate the real trajectory correctly.

Abstract：According to the characteristics of industrial control system, based on the analysis of the existing industrial control system programming standard IEC61131-3 stipulated industrial language, this paper studies the formal verification method based on industrial language, analyzes the ST and LD languages, and gets the finite state model of the machine to achieve accurate description of the control objectives. The finite state machine model is verified by NuSMV, and the result of formal verification is gotten. In this way, the PLC logic code of IEC61131-3 programming language is analyzed, the formal verification model is established, and the possible logic defects of PLC logic code written by users are found, and the report of analysis and verification of these defects is provided.

Abstract：In this paper, an active intrusion tolerance control strategy is designed for the inevitable false data injection attack in the control system of island AC microgrid when there is external interference. Firstly, the state space model of the island AC microgrid inverter system is constructed. A sliding-mode attack observer is designed to estimate both the state variables and the false data injection attack information in the control system. After obtaining the estimated information of the attack, the integrated sliding mode intrusion tolerance controller is used to actively control the false data injection attack and external finite energy interference to ensure the stable operation of the microgrid system. Finally, the feasibility and effectiveness of the proposed method are verified by simulation.

Abstract：In order to solve the long-term pain point that the household appliance industry relies on manual appearance quality inspection, this paper innovatively applies the key technology of 5G + MEC machine vision, and constructs an intelligent product surface quality inspection system. The detection system operates in the cloud side collaboration mode, supports edge reasoning of pictures and videos download and deployment of cloud algorithm terminals, supports heterogeneous computing, and has high real-time performance. The system comprehensively manages and applies the collected data based on decision analysis tools, artificial intelligence, big data analysis and other technologies. At the same time, the system is associated with MES, SCADA, WMS and other systems of enterprises，and integrates production end data, material data and market tracking feedback data，which can realize data interconnection and quality traceability in the whole life cycle of products.

Abstract：This paper analyzes the current characteristic curves of photovoltaic modules under different working conditions and finds that the current data of photovoltaic modules superpose complex performance characteristics and high noise. In order to accurately diagnose the fault types of photovoltaic modules, a soft thresholding temporal convolutional network(ST-TCN) photovoltaic module fault diagnosis model is proposed. The ST-TCN network uses the dilated convolution layer, ReLU layer, and Dropout layer of multiple residual modules to extract current numerical and time series features, uses the soft thresholding of residual modules to de-noise the extracted features, and finally uses the full connection layer to diagnose and classify the extracted features of residual modules. The experimental results show that the ST-TCN network has a simple structure, fast convergence, and high accuracy in fault diagnosis, reaching 92.99%.

Abstract：In this paper, the H∞ filter problem under external disturbance is studied for the inductive coupled power transfer(ICPT) system with the series-parallel(SP) resonant compensation network. To suppress external disturbance and obtain ideal estimation results at the same time, a generalized state space equation based on the circuit topology of the SP-ICPT system is established firstly. Then, a series of H∞ filtering gains is obtained by the asymptotic stability judgment condition of the filter error system. Finally, a simulation with regard of SP-ICPT system is carried out to verify the accuracy and feasibility of the theory proposed in this paper.

Abstract：Applied on machine learning, malicious web page detection technology is studied in this paper. At present, popular methods of “feature code” or “whitelist” can only detect known malicious web pages. The method of machine learning can detect unknown malicious web pages, but it has to face the problem of that the data is large, complex and tedious when processing web page features. In this paper, a Hash compression method is proposed. The method can map 1.5 million features into 20,000 feature space, and train multiple traditional machine learning models and integrated learning models using k-fold cross-validation method for extracted feature data. The best classification detection model will be selected by evaluating the detection effect of the model.

Abstract：With the rapid growth of demand for computing power in various fields of society, more and more attention has been paid to improving computing power of computing systems through computing architecture innovation. The development of computing system is a complex system engineering. Innovations in computing architectures have resulted in ever-increasing complexity of computing systems. Traditional systems engineering methods are difficult to meet the needs of system development. The modeling method of a kind of computing system is studied by the method of model-based systems engineering(MBSE), the modeling framework is proposed, and the system model is established. Using this modeling method, a model of a localized computing system was established, which proved the effectiveness of the modeling method, and laid the foundation for the next step to use computer simulation technology to develop a type of computing system.

Abstract：To promote independent innovation ability in the information technology field, it is necessary to develop high-quality computer products by conducting the joint research on computer software and hardware via systematic method. This paper proposes a novel model to deal with the current challenges in the joint research on computer software and hardware based on PKS system, which includes the innovations on the aspects of management, service and communication. This model has been applied in some real projects and has achievd positive results, which is valuable and instructive for the future work.