This PDF file contains the front matter associated with SPIE Proceedings Volume 12987, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

With the continuous development of modern communication technologies, many computation-intensive and delay-sensitive applications have emerged. Unmanned Aerial Vehicle (UAV) can perform excellent and unique in emergency communication systems by virtue of its flexibility. In this paper, a UAV-assisted fog computing emergency communication system is designed, wherein a UAV that has computing capabilities can act as a UAV Fog Access Point (UAV-FAP) and offer surrounding mobile devices (MDs) computational offloading services. The objective is to minimize the maximum processing delay of the system. The optimization problem of jointly optimizing MD scheduling, distribution of resources and UAV flight direction is proposed under energy consumption constraints. Considering that the problem is a mixed integer nonlinear programming (MINLP) problem, a computational offloading algorithm based on Deep Deterministic Policy Gradient (DDPG) is proposed to solve the non-convex problem. And the Switches from Adam to SGD (SWATS) algorithm is also invoked to update the neural network parameters with the optimizer to avoid falling into local optima to some extent. Simulation results show that the DDPG-SWATS joint algorithm converges quickly and has a small processing delay compared with the benchmark algorithm.

This paper presents the design of a 12-bit 4 MS/s successive approximation register (SAR) analog-to-digital converter (ADC) based on the 40nm CMOS process. The internal structure of the ADC employs an "8MSB+4LSB" hybrid capacitive-resistive structure in the digital-to-analog converter (DAC) design to reduce layout area. The comparator circuit utilizes a dynamic comparator structure consisting of a cascade of three pre-amplifiers and a single latch, with output offset storage used to minimize comparator offset voltage. The layout design is symmetric to enhance device matching. In post-simulation under the conditions of 1.1 V analog voltage, a 941.40625 KHz input sine wave signal, and a 4 MHz sampling clock, the SAR ADC achieved an effective number of bits (ENOB) is 11.88 bits, a signal to noise and distortion ratio (SNDR) is 73.2941 dB, a spurious-free dynamic range (SFDR) is 89.4 dB, a power consumption is 1.18 mW, and a figure-of-merit (FoM) is 168.6 dB.

With the widespread use of wearable devices, human activity recognition (HAR) holds immense potential in health monitoring, smart environment. Notably, temporal sensory sequences collected from the wearable devices can provide accurate reflections of the daily activities. Nonetheless, existing CNN-based and LSTM-based methods have predominantly concentrated on feature extraction from univariate sequences, overlooking the implicit frequency information. Therefore, we firstly employed the Short Time Fourier Transform (STFT) in HAR tasks, extracting inherent frequency feature. Concurrently, we introduced a multi-branch network that combines CNN and LSTM. The CNN component captures spatial information of different dimensions. The LSTM, on the other hand, comprises two parts, one focused on temporal relationships within a single channel and the other concerned about channel relationships at a specific time point. In addition, recognizing the limitations in the available datasets, particularly the insufficient coverage of daily activities, we collected our custom dataset, encompassing eight distinct daily activity categories. Finally, we evaluated our proposed model and benchmark models. The results demonstrate that our network exhibits superior generalization across different datasets, achieving accuracy of 91.70%, 95.79%, 87.81% on the PAMAP2, UCI HAR and our own dataset respectively.

With the increasing scale and complexity of the power system, timely monitoring and accurate identification of abnormal data in power grid reports are crucial for ensuring the safe operation of the power grid. This paper proposes an automated alarm method for abnormal data in power grid reports based on the Paxos algorithm. Automatically collect power grid report data using RPA technology, and then use the isolated forest algorithm of sub forest progressive update incremental learning to detect anomalies in the collected data. Based on the anomaly detection results, the Paxos algorithm is used to trigger the automated alarm system and promptly notify the operation and maintenance personnel. The experimental results show that this method can accurately detect abnormal situations in the power grid and improve the comprehensiveness of alarms, which is beneficial for improving the security of the power grid.

To make the ion thruster thrust can be continuously adjusted to the set thrust value and meet the adjustment accuracy, it is necessary to study the ion thruster thrust adjustment method. The traditional thrust adjustment method has narrow thrust adjustment range and single adjustment point, which can’t meet the requirements of continuous thrust adjustment. Based on the incremental PID closed-loop control algorithm, the control principle of the algorithm was analyzed, and the PID parameters were set through experiments to get the calculation method of thrust closed-loop control. The algorithm is implemented by software to adjust the output of ion thrusters, the results show that the proportional thrust control accuracy of the ion thruster reaches 0.45%, the minimum thrust control accuracy is 99.2% in the stationary section, and the thrust closed-loop control accuracy is ±0.1 mN. The algorithm can quickly control the ion thruster and make its thrust value reach the target value. Moreover, the ion thruster has continuous adjustable thrust, wide adjustment range and stable thrust output, which can be well applied to the propulsion system of ultra-low orbit Earth observation satellite.

Passive source localization based on time difference of arrival (TDOA) is an important method for high-precision and fast localization of point sources in areas such as gunshot localization and remote missile landing point localization. Traditional multi-element array passive sound localization algorithms have nonlinear equation, and their analytical solutions are obtained through complex transformations, which may result in non-unique solutions and low localization accuracy, making them unsuitable for practical applications. In this paper, a five-element tetrahedral array localization model is proposed, which utilizes the redundant degrees of freedom contained in the five-element array and introduces an intermediate variable (redundant variable) to linearize the resulting equation system. The accuracy of the analytical solution obtained by this optimization algorithm and the directional and distance errors under this array model are analyzed.

In modern communication systems, satellite networks play a crucial role in providing global coverage and high-speed data transmission. Traditional network simulation tools face challenges such as high complexity, poor scalability, and inadequate performance, making them unsuitable for simulating satellite networks. To gain a better understanding of satellite network performance and operation, this paper introduces a Docker-based satellite network simulation platform. This platform simulates satellite constellation link establishment rules and uses Docker to construct satellite network topologies, automating snapshot switching, link simulation, satellite communication, and other functions. The platform provides an experimental environment for the design, optimization, and performance evaluation of satellite communication systems, aiding researchers in gaining deeper insights into the operation and performance characteristics of satellite networks.

Webshell is a type of malicious script that is written by network attackers based on the characteristics of web programming languages. It usually exists in the form of web files such as ASP, PHP, JSP, or CGI. As a type of website backdoor, attackers often mix ASP or PHP backdoor files with normal web files in the web directory of the website server, and then use a browser to access the ASP or PHP backdoor in order to control the website server. With the rapid development of computer technology, the importance of web security has become more prominent. This article starts from the current mainstream webshell detection methods, analyzes the mechanism and characteristics of webshell detection, and proposes and implements a webshell detection model based on ensemble learning combined with machine learning. The test results show that when the proposed detection method is combined with the feature extraction method, the recall rate reaches 89% and the accuracy rate reaches 95%, which basically meets the requirements of recall rate and accuracy rate in emergency response scenarios for network security.

Autonomous vehicles are the core of future smart transportation, and Tesla is a leader in the field. As vehicles become intelligent, cyber attack defense becomes a concern. Tesla autonomous vehicles may be challenged by Stuxnet-style malware, whose self-replication and propagation capabilities may threaten transportation security. This study proposes a spatiotemporal mathematical model for such malware. Based on computer virus propagation dynamics, it is combined with Tesla Autonomous vehicles for the first time. The propagation characteristics of the malware are analyzed, and scientific recommendations are provided for future prevention.

Aiming at the problem of low radar emitter recognition rate in electronic reconnaissance under low signal-to-noise ratio (SNR) conditions, a radar emitter recognition method based on kernel extreme learning machine optimized by sparrow search algorithm is proposed. By constructing a radar database, the kernel extreme learning machine optimized by sparrow search algorithm is used to identify radar data under different noise conditions, and is compared with the traditional kernel extreme learning machine and support vector machine method. The results show that the recognition accuracy of the proposed method is higher than that of other methods in different noise conditions.

The analysis of inertial confinement fusion (ICF) plasma diagnostic pictures is crucial for fusion energy research. In this paper, a method based on the combination of deep reinforcement learning and computer vision techniques is proposed for the analysis of ICF plasma diagnostic pictures. The method first preprocesses the images using computer vision techniques and then uses deep reinforcement learning to classify and recognise them. The physical quantities are closer to the theoretical values using the new method, which is more instructive for the experiments. For example, at the radiation temperature, the obtained values are increased by 20-70 eV, and at the electron and plasma temperatures are close to the theoretical 5 KeV. at the same time the neutron yield is increased by a factor of 10. The experimental results show that the method has high accuracy and efficiency in the analysis of ICF plasma diagnostic pictures, and can effectively assist fusion research.

To solve the internal attack problem of the wireless sensor network, this paper proposes a credibility-based trust model by using the existing classical credibility mechanism. The model consists of three levels: credibility calculation, trust management and trust decision-making. Credibility calculation is the core of this model. The direct credibility, indirect credibility and energy credibility are combined to form a comprehensive credibility. The network energy consumption is reduced by updating the credibility regularly. The credibility threshold is set to make trust decisions, and effectively distinguish malicious nodes and aging nodes, thus to ensure the security and reliability of the network. The simulation results show that the model can detect malicious nodes effectively, has a high detection rate, and greatly reduces the energy consumption of the whole network.

In the current social environment, UAV detection technology has various applications in various fields, such as agricultural management, urban planning, and security monitoring. However, there are still problems, such as difficult localization and low accuracy in the UAV small target detection field. In this paper, for the problem of UAV small target detection, the characteristics of YOLOv7 model are sorted out and improved from several aspects, and the AED-YOLOv7 model is constructed. Aiming at the problem that multi-scale feature fusion in the model is prone to feature loss, this paper uses the FAM module, which reduces the spatial location variability of information between multiple scales and strengthens the feature fusion capability. This paper introduces the EMA module, which learns effective channel descriptions, generates better pixel-level attention for advanced feature mapping, and groups channel dimensions into multiple sub-features so that spatial semantic features are evenly distributed in each feature group. In this paper, we use Decoupled Head to process the classification and regression tasks separately, which can be better adapted to the needs of different tasks. Finally, this paper uses the EIoU Loss function to minimize the difference between the width and height of the target frame and the Anchor while retaining the beneficial properties of the CIoU Loss, which produces faster convergence and better localization results. On the VisDrone2019 benchmark, compared to the performance of YOLOv7, the AP₅₀ of AEDYOLOv7 is improved by 3.5% and APS by 3.2%.

Aim at the low efficiency of the traditional incremental conductance method and the insufficient adaptability of the orthogonal vector construction of the single phase locked loop, this paper introduces the incremental conductance method combined with the constant voltage method to achieve the maximum power point tracking (MPPT), and uses second-order generalized integrator based phase locked loop (SOGI-PLL) to construct the orthogonal components to achieve the phase locked. In Matlab/Simulink, a simulation model of the single-phase photovoltaic energy storage grid-connected inverter is constructed and simulated. The simulation results show that not only the bus voltage is stable, but also the grid-connected current is in phase with the grid voltage, and the value of THD meets the grid-connected requirements, which proves the effectiveness of the above control method.

With the rapid development of civil UAV technology, their wide application in various fields such as agriculture, transportation, logistics and monitoring has become reality. However, to enable safe and reliable UAVs, as well as longtime operation in special environments, it becomes crucial to design an effective operation and maintenance control system for UAVs. Our team proposes a civil UAV intelligent operation and maintenance system based on multi-source sensor data, which realizes real-time flight status and flight data monitoring and reality of UAVs by integrating LiDAR, vision system and other sensors, which can automatically enter the hangar for charging and maintenance according to the instruction, etc. Finally, a comparative test was conducted on some of the performance of the unmanned aerial vehicle system, demonstrating the effectiveness of the relevant algorithms of our system.

With the vigorous development of big data technology, the application of data mining program to complete quantitative prediction and evaluation of mineral resources has become a new trend in the development of mineral exploration industry and an important way to promote the digital transformation of traditional prospecting technology. In this regard, this paper will combine geological theory analysis and data mining algorithm to build a quantitative prediction model of metal mineral resources, realize the analysis, processing and visual interpretation of deep geoscience data, and provide necessary data science support for actual mineral exploration. The prediction model takes the machine learning algorithm as the core, and relies on the class libraries such as Numpy, Pandas and Sklearn in Python environment to complete the development and training. Combined with the spatial attribute database in ArcGIS, the prediction elements and importance evaluation indicators are obtained, and the final quantitative prediction results are obtained, and the result evaluation is completed. Practice has proved that the quantitative prediction model of mineral resources based on machine learning algorithm has high prediction accuracy, and the visual results can clearly show the high probability mineral resources areas, which has high practical promotion value.

For the detection of underwater metal objects, electromagnetic detection technology is not limited by hydrological conditions, and has the advantages of stability, reliability, strong concealment, strong recognition ability and high positioning accuracy. Active electromagnetic detection achieves the detection of metal objects by radiating electromagnetic fields and detecting the induction characteristics of the target. This paper designs an active electromagnetic detection method for underwater metal targets based on the adaptive compensation scheme of active electromagnetic primary field, and builds an active electromagnetic detection system, which successfully detects cylindrical targets with a diameter of more than 150mm within a distance of 4m.

In order to meet the needs of Internet of Things devices and wearable electronic devices for lower standby power consumption and power supply noise, a Low Dropout Regulator (LDO) circuit with low quiescent current and high power supply rejection ratio is designed. The core circuit is bandgap and error amplifier. This design is based on the 0.18um BCD process. The Spectre simulation of Cadence has been used to test and simulate the circuit. The simulation result indicates that the bandgap voltage is 1.209V, the temperature coefficient is 19.76ppm/°C in the range of -25~125°C, and the power supply rejection ratio is 69dB at low frequency. The circuit can deliver a quiescent current of around 776nA and exhibits good temperature characteristics and low quiescent current.

The traditional substation location of airport peripheral power grid only considers the cost and income of energy storage equipment, but does not fully consider the civil aviation clearance limit surface. Therefore, a particle swarm optimization method for substation location of airport peripheral power grid considering the civil aviation clearance limit surface is proposed. The civil aviation clearance limit surface is constructed, and the load planning structure of power grid substation is established. The location objectives are set as the minimum investment, the lowest loss and the highest voltage margin, and the location objective function based on particle swarm optimization method is calculated to realize the particle swarm optimization of power grid substation location around the airport. The experimental results show that this method improves the location problem of substation around the airport, improves the convergence speed of the method, reduces the time to obtain the best value, and reduces the loss rate of the optimized power grid.

With the development of information technology and electric power industry, modern power grid system gradually integrates computer intelligent technology. Distributed photovoltaic network, as an important power supply system using new sustainable energy, contains a large number of intelligent terminal equipment, which leads to the distributed photovoltaic network facing more serious security threats. In order to promote the construction of distributed photovoltaic power network security protection system, this paper proposes an intelligent device fingerprint access technology for distributed photovoltaic power network system. Based on adaptive filtering, the device anti-forgery fingerprint is generated, which involves the field of artificial intelligence and network security. Among them, through the innovative combination of the adaptive filtering algorithm NLMS and LMS algorithm, the weighted average optimization of the feature set, combined with the artificial intelligence classification and recognition algorithm, the specific characteristics of the parameters at the convergence of the adaptive filtering technology are achieved, and efficient device fingerprint generation is realized. It can realize the device authentication of the distributed photovoltaic network intelligent terminal equipment access, so as to effectively improve the comprehensive defense capability of the distributed photovoltaic network.

With the development of high-power and high integrated radio frequency (RF) electronics, the thermal issue of high heat flux chip cooling is becoming urgent. Liquid cooling is preferable for RF electronic heat dissipation. Aluminum cold plate with single layer series flow channel is widely used for current RF electronics. However, the increased convective temperature rise and pressure drop restricts its application in future high heat flux RF electronic cooling. In this work, aimed at improving high-power RF linear chip array cooling ability, we propose a two-layer aluminum cold plate composed of liquid dividing layer and heat dissipation layer, forming a fully-parallel microchannel structure. The flow and heat transfer performances of the single-layer and two-layer cold plate are comparatively evaluated using a finite-element numerical simulation. Then, the effects of key parameters of the two-layer cold plate are analyzed, and the guidelines for choosing the parameters are presented. The results indicate that the proposed two-layer cold plate can simultaneously reduce the convective temperature rise, pressure drop and chip temperature deviation by 48%, 65%, and 67%, respectively.

An operational amplifier based on UMC 40nm process is designed. PMOS and NMOS differential pairs are adopted in the input stage of the circuit to achieve rail-to-rail input, and the minimum current selection circuit is used to reduce the change rate of the transconductance of the input stage. The gain stage adopts a common source common gate structure. In order to obtain full swing output, class AB output stages are adopted, and the output MOS is biased by feedback technology. The CMRR of 85.14 dB and PSRR of 85.86 dB can be obtained. The open-loop gain is 90 dB, the phase margin is about 83°, the unit-gain bandwidth product is 15.84MHz, and the offset voltage is 283.9nV.

The principal objective of this project was to solve the problem of the controlling attitude balance for the cube to achieve actuator saturation. For the past three decades, studies of attitude balance for the cube have ignored the impact of actuator constraint. Firstly, a novel non-linear saturation function is constructed that has the characteristics of being continuous and smooth. Secondly, based on the proposed saturation function, a non-linear proportional-derivative (PD) saturated control (NPDSC) is developed to tackle the attitude balance problem for cube formulated by Lagrange equation. It is proven that the designed NPDSC can assure that the system of cube is globally asymptotically stabilized. The control method offers the benefits of a straightforward and concise format, efficient calculations, and the absence of any specific numerical value associated with the system. Another feature is that the proposed regulation guarantees the output of the output generator is maintained within the rated range, so it completely avoids instability and degradation or unimaginable movement of the system, avoiding excessive output due to overheating or structural failure. The proposed control method actually provides an efficient solution for cube systems with output saturation. The simulation results provide evidence of the feasibility of this approach.

In this paper, a transistor with Si as the substrate, SiO₂ as the dielectric layer, and Pentacene as the organic layer, and Au as the source-drain electrode was designed and prepared, and the dielectric layer was optimized by spin-coating a layer of PMMA. Subsequently, the transistors pre and post optimization were characterized by Raman detection, white light interferometric detection, XRD and other characterization tests, which verified that the optimized Pentacene film has better properties. Finally, electrical measurements of the two devices were carried out using an electrochemical platform. The output versus transfer characteristic curves derived from the tests confirmed that the PMMA-optimized parallel Pentacene thin-film transistors had enhanced mobility, boost significantly lower threshold voltage, and reformed device performance.

This article designs a dual-band microstrip patch antenna in the 12-18 GHz range. The antenna works in the 12.80-13.10 GHz and 16.26-17.05 GHz frequency bands, respectively. When the antenna works at low frequency, the return loss is less than 10 dB, and the maximum gain is 4.7 dB; when the antenna works at high frequency, the return loss of the antenna is less than 10 dB, and the maximum gain is 6.4 dB. At the same time, the antenna has good radiation characteristics within the working frequency band and can meet the requirements of wireless communication.

Lightweight composites with flexibility and ideal electromagnetic interference (EMI) shielding performance are required for protecting electronic devices against electromagnetic wave pollution. Herein, binary flexible and conductive polyvinyl alcohol (PVA)/Ti₃C₂T_x composite film with a “brick-and-mortar” structure was prepared using a vacuum filtration self-assembly strategy. By the combination of PVA and Ti₃C₂T_x, the PVA/Ti₃C₂T_x composite films have been successfully obtained, leading to high electrical conductivity (≥12.3 S/cm) and superior EMI shielding efficiency (≥27 dB, 8.2-12.4 GHz). The preparation of PVA/Ti₃C₂T_x film offers a promising approach for designing and fabricating flexible high-performance MXene-based EMI shielding materials, which may be applied in various applications including flexible wearable electronic devices and portable equipment.