Abstract：As modern life becomes increasingly intelligent, more and more applications require inferring semantic information from images before proceeding with further processing, such as virtual reality, autonomous driving, and video surveillance. Current semantic segmentation models achieve ideal performance through supervised training with a large amount of annotated data, but their performance severely deteriorates when inferring on data with a distribution different from the training data. This means that once the application scenario changes, new data needs to be annotated and the model needs to be retrained with the new data in order to achieve normal performance. This is undoubtedly time-consuming and expensive. Therefore, domain adaptive semantic segmentation algorithms provide a solution to the problem of the model's performance degradation on data with different distributions. This article summarizes the cutting-edge progress of domain adaptive semantic segmentation algorithms and looks forward to future research directions.

Abstract：Standardization plays a fundamental and leading role in supporting industrial development and promoting progress of science and technology. It also provides a new perspective for research in industry development. Based on domestic and international achievements and advancements in quantum secure communication standards and combined with the practical applications of major infrastructure, applications, and the demands of next-generation ICT technologies, this paper explores and analyzes the characteristics of the integration of quantum communication and the ICT system and observes that the integration of quantum communication with key ICT technologies such as 5G, cloud platforms, and the IoT is an important trend for future development. Particularly, as the integration of quantum communication and ICT technologies is progressing ahead of the developments of standardization, there is an urgent need to optimize the layouts of standardization, promote relevant scenario applications, and facilitate the development of quantum communication as a future industry.

Abstract：Garbage classification is an important part of building ecological civilization. To solve the problem that heavyweight models are difficult to deploy to mobile devices, an improved garbage image classification method based on YOLOv5 network is proposed. The backbone network fused with GhostNet is used to replace the traditional convolutional operation with linear operation, which reduces the number of parameters of the model and improves the model inference speed. By adding an improved version of channel attention module to the network, the important channel features are strengthened and more deep-level feature information is obtained. The weighted boundary fusion method is used to improve the localization accuracy of the detection frame. It is experimentally demonstrated that the method improves the accuracy by 8.5%, reduces the parameter quantity by 46.7%, and increases the average inference speed by 1.22 ms in the homemade dataset compared with the original model, achieving a comprehensive improvement of accuracy and inference speed.

Abstract：Once a task-based question answering system is built, it is usually fixed and can answer very limited questions, making it difficult to meet user needs. A method for automatically updating the knowledge base in real-time was proposed. When a user asks a question that the question answering system cannot answer, the system will automatically send the question to the manual customer service. After the manual customer service used professional knowledge to reply, the system can automatically obtain the user's question and the answer replied by the manual customer service in real time, and automatically update the question answering pair to the knowledge base in real time. If other users ask similar questions, the question answering system can quickly provide corresponding to answers. Taking the question answering system in the field of government affairs as an example, the text vectorization method ERNIE was applied to build a question answering system that automatically updates the knowledge base in real time. After computer experiments, it has been proven that the proposed method can achieve automatic real-time updates of the knowledge base, and the constructed question answering system has autonomous learning and memory functions, improving the intelligence level of the task-based question answering system.

Abstract：In this paper, a new type of over-temperature protection (OTP) circuit for LED driver is designed. It is based on 0.18μm BCD process. In this paper, Banba′s bandgap reference based on current summing is used to generate high and low threshold voltage, to eliminate the oscillation phenomenon near the over-temperature point of the chip. The bandgap reference is added with high-order curvature compensation to improve the accuracy of the over-temperature threshold voltage. The circuit is simulated by Cadence software. The results show that the high and low threshold voltage has good temperature property, the temperature coefficient is 2.9 ppm in a temperature range of -40 ℃ to 150 ℃. When the temperature is higher than 131.8 ℃, OTP can be triggered. When the temperature is lower than 109.4 ℃, the circuit can return to normal operation， and the hysteresis quantity is 22.4 ℃. The OTP has high precision and stability, and can be applied to LED driver chips.

Abstract：With the continuous growth of IC design scale and complexity, traditional EDA verification is more and more restricted especially for sub-system and SoC full chip level simulation efficiency. To efficiently and effectively find and debug RTL problems and ensure the IC delivery quality and time to market has become the critical path for IC development. This article introduces a platform (DVA) based on distributed computing method to accelerate large scale IC verification. The DVA platform based on UVM, adopts native Linux socket components as underlying layer for communication.

Abstract：For Controller Area Network(CAN) controller with asynchronous interface, a randomized and reusable function verification platform based on Universal Verification Methodology(UVM) is designed and implemented. The platform is built using object-oriented UVM classes. The code is more reusable and the development time is shorter. The introduction of a more randomized incentive accelerates the convergence speed of functional verification, and the random incentive is closer to the actual application of the chip. The automatic comparison mechanism can output the result report in real time, which is convenient for problem debugging.

Abstract：Through the study and understanding of the film bus, in order to maximize the advantages of the bus, this paper combines the direct storage and access technology and ICB bus, and designs an ICB bus based DMA controller. The logic function of the controller is verified by software simulation. The simulation results show that it can run normally and stably.

Abstract：In order to meet the requirements of online electrical performance parameter testing for various small electrical components of metro vehicles under the accelerated test， an online test platform for small electrical equipment of metro vehicles based on accelerated test was proposed through the design and improvement of the traditional online test power supply， test equipment and data processing system. According to the special requirements of small electrical equipment for metro vehicles， the design idea of power supply with AC and DC， adjustable voltage， high power and high precision was put forward. The test link of electrical signal acquisition and transmission was designed， and the feasibility of the test link was proved in the actual experiment. An online test system which can process the electrical signal data in real time and calculate and store the electrical performance parameters of small electrical equipment of metro vehicles was constructed.

Abstract：In order to realize the waveform reconstruction of high-frequency signals under under-sampling conditions, the generation of progressive delay ultra-narrow trigger signals has become the key to the use of sequential equivalent sampling techniques for high-frequency signal sampling in trigger sampling systems. To this end, a programmable delay trigger sequence generation and conditioning circuit is designed. Under the control of the FPGA digital circuit, an adjustable delay trigger sequence is generated through the counter and delay module, and the step recovery diode characteristic is utilized to condition the generated trigger sequence to produce a pulse signal with adjustable delay step and extremely narrow edge. By testing the circuit, the results show that the output pulse signal step range is adjustable between 0~2.4 ns, the resolution can be up to 1 ps, the edge jump time can reach within 120 ps, and the amplitude is about 8 V. The delayed pulse and conditioning circuit can be applied to communications, radar and other signal detection equipment, and has great significance for the acquisition and analysis of high-frequency signals.

Abstract：In order to solve the problem that the traditional energy detection algorithm has poor detection performance due to the fixed decision threshold when the signal-to-noise ratio is low, this paper studies a dual-threshold cooperative energy detection algorithm that can change with the signal-to-noise ratio coefficient. The threshold value of the algorithm is adaptively adjusted according to the signal-to-noise ratio coefficient. At the same time, considering that the centralized cooperation is incompetent due to the problem of over-reliance on the fusion center, the distributed cooperation method is used to integrate the information detected by each cognitive user. The energy value is compared with a dynamic threshold. Finally, it is proved that the algorithm has better detection performance comparing with the two traditional algorithms.

Abstract：In actual spectrum sensing scenarios, the primary user may arrive or leave randomly, and when the primary user state changes dynamically during real-time spectrum sensing, the performance of the existing static spectrum sensing algorithm deteriorates sharply. For this situation, this paper propose a dynamic primary user spectrum sensing algorithm based on the residual shrinkage and attention mechanism. During the spectrum-sensing interval, the time when the primary user randomly arrives or leaves randomly follows a uniform distribution. The “deep residual shrinkage network (DRSN)” is used to extract dynamic primary user features and filter out redundant noise features. The “coordination attention module (CAM)” is used to improve the ability of the model to express the features of the dynamic primary user. Simulation results show that the proposed algorithm performs are better than ResNet algorithm, CBAM_IQ algorithm and CBAM_Energy algorithm. The proposed algorithm can maintain a high detection probability for the primary users who randomly arrive or leave following different distributions.

Abstract：To address the high cost of implementing long-bits RSA encryption algorithms in hardware, improvements have been made to the traditional radix-4 Montgomery algorithm. Firstly, a Carry-Save Adder (CSA) is introduced to efficiently complete large numbers of addition computations. Optimization is then carried out in post-processing to reduce the number of large numbers to be calculated at each Montgomery computation. Finally, a pipelining technique is incorporated into the RSA encryption algorithm for parallel execution, thereby reducing the use of hardware resources. Experimental results on a Xilinx XC7K410T FPGA development board demonstrate that while maintaining encryption speed, the hardware resources used by the improved RSA encryption algorithm are only half of those of the previous parallel structure and can operate at higher frequencies.

Abstract：To meet the requirements of “measurement in motion” for narrow beam TT&C equipment in the vehicle/ship state, a method of “tracking loop & feed-forward compensation of ship swing & gyro stabilization loop” for vehicle/ship swing isolation is proposed. Through simulation and annalysis the ship swing isolation and vehicle swing isolaiton, the simulation results show that with this method, the ship swing isolation can reach above 50 dB and the vehicle swing isolation can reach above 20 dB; The method is used to simulate and analyze the tracking performance of high-speed moving targets under the conditon of ship swing, the simulation results show that the tracking error meets the tracking requirements of narrow beam TT&C equipment.

Abstract：In order to solve the controlling difficulty of the diversified heterogeneous resources of industrial network cyber range, a various-resource-based meta-modeling approach is adopted to model, classify and describe the functions, performance, interfaces, forms and other attributes of the real resource in a unified way, which maps the real resource to achieve model reuse and rapid reconstruction. Based on analyzing the management and control interface, collection interface and I/O interface of the resources, this paper establishes and realizes the component based resource encapsulation specification, so as to provide technical support for the flexible allocation and automatic management of industrial control network cyber range resources. The method of resource modeling encapsulation is verified by software tools. The significance of this paper is to realize the visual operation and unified management of the resources through modeling encapsulation technology.

Abstract：In this paper, according to the needs of airborn ultra-short wave receiver, the system-level design and simulation of RF front-end circuit is carried out by using Advanced Design System（ADS）. By analyzing the performance requirements of ultra-short wave receiver, combined with the main working principle and technical indicators of the receiver, the structure of the dual-conversion superheterodyne receiver is selected as the implementation scheme. The system level simulation model of RF front-end of receiver is established in ADS. The key indicators of RF front-end of the design such as noise figure, sensitivity, gain and intermodulation distortion are designed, analyzed and simulated by ADS. The simulation results show that the RF front-end meets the requirements of the design.

Abstract：The main signal source is one of the key components of the synchrotron radiation device. It is not only used to generate the highly stable reference signals required by each subsystem of the synchrotron radiation light source, but also used to generate the high-precision working clock required by the control system of the entire device. Generally, one RF signal source is used as the main signal source. But the commercial RF signal source generally has only one output channel, which is far from meeting the needs of the synchrotron radiation device, while the use of traditional power splitters to expand the main signal channel also suffers from amplitude attenuation, accuracy degradation, and inconsistent phase. In order to solve the above problem, the method of extending one channel output of microwave signal source with RF shortcut chip AD9361 has been proposed in the paper.

Abstract：Aiming at the real-time switching and flexible intelligent invocation requirements of various life services supported by satellites, a scheme for real-time, efficient and dynamic switching and deployment of processor functions is designed based on the heterogeneous processor of wireless wide-area signal services. Through the customization of multiple functions of large-resource FPGA and multi-chip 8-core DSP combined with dynamic deployment design, real-time dynamic reconfigurable processor system functions are realized, and five types of FPGA are combined with two types of DSP applications to dynamically combine control, data links, and complete multi-task intelligent switching.

Abstract：In this paper, a wireless implantable optogenetic hardware structure is proposed. In the experimental process, a magnetic resonant wireless energy technology is used outside the organism to power the in vivo circuit; however, in the Wireless Energy Transfer (WPT) process, it is necessary to convert the electrical energy at the transmitting end into a high-frequency magnetic field with the help of magnetic coupling, and the magnetic field is the medium for transmitting electrical energy in the wireless energy transfer system, and the Electromagnetic Radiation (EMI) problem brought by it will be a threat to the organism. Therefore, an improved frequency spreading and harmonic suppression wireless energy supply circuit is designed.

Abstract：The portable energy dispersion XRF spectrometer is powered by Li-ion battery. The output voltage of Li-ion battery gradually decreases with the stored energy is released, which causes a bad impact to the detection results of the spectrometer. To solve the problem, a voltage stable converter based on LTC3789 and Buck-Boost topology is designed which has a wide range input voltage and high efficiency. The design process of the converter is discussed in detail and the function and efficiency are tested. It is proved that the converter has stable output voltage, high efficiency and practicability.

Abstract：The current indoor and outdoor joint positioning technology based on GPS and WiFi achieves high positioning accuracy. However, this type of method needs to process signals of two different scenarios in the transition area, and there are problems such as difficulty in smooth switching between areas, and lack of effective data association. In response to the above problems, this paper associates the physical space location information with WiFi information to establish an indoor and outdoor joint positioning map model based on WiFi signals, and uses the Jaccard similarity algorithm for indoor and outdoor judgment and one-time positioning. In order to increase the positioning accuracy of WiFi, this paper uses cubic spline interpolation to expand the WiFi fingerprint library, and uses the weighted K-nearest neighbor algorithm for secondary precise positioning.

Abstract：In family and medical health monitoring, timely and accurate real-time monitoring of elderly falls is an important topic, and millimeter wave radar has shown great application potential in non-contact fall detection. The traditional method mainly uses the Doppler information obtained by millimeter wave radar as the fall detection feature. However, for some scenarios, such as the distance between the person and the radar or the direction of the fall is perpendicular to the normal of the radar antenna, the effectiveness of the Doppler information feature will be reduced, and the fall detection accuracy will be reduced.