This PDF file contains the front matter associated with SPIE Proceedings Volume 13401, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Ultra-high-speed permanent magnet motors are widely used in turbine machinery due to their high speed, high power density and small size. A 5kW/150000RPM ultra-high-speed permanent magnet motor which is used for vehicle electric turbo-compound system was analyzed, mainly focused on rotor strength, motor loss and temperature rise. First, based on the classical motor design theory, the design steps are defined and the electromagnetic design is completed; then, a 2D simplified model of the rotor is establish and the rotor strength is analyzed through multi-physics coupling method; finally, based on the motor loss theory, the thermal characteristics of the motor are analyzed; the result shows that the designed motor can satisfy the requirements, and provides references for motor development and optimization design in the future.

This paper introduces the research on automatic drilling technology of robot for analog products around the typical components of the access door. Based on the existing assembly conditions of the access door production line, a set of analog product robot automatic drilling system scheme has been formed through seven key steps: AGV drilling robot cup cone design, Modeling of the Envelope Shape of Analog Fixtures, Design of Automatic drilling Pre-Assembly Process, Robotics - Product position and posture establishment, Revision and Verification of Fastener Positions for Analog Products, tooling design, and drilling Parameter Experimentation and Adjustment. This reduces the labor intensity of workers and improves the position accuracy and quality accuracy of connecting holes.

The control problem of nonlinear network switching system with time delay has been studied. In the presence of system uncertainty, the T-S model is employed to represent the nonlinear network switching system as a network switching fuzzy system. The H∞ robust control problem of the system is studied when there is interference in the network. The multiple Lyapunov function method is employed to design a state feedback controller and a switching signal, so that the network-based switched fuzzy system has H∞ performance index γ. A sufficient condition is provided to address the H∞ robust control problem for these systems.

In order to solve the shortcomings of the traditional constant step incremental conductance method in both speed and accuracy, an improved variable step incremental conductance method is proposed. The simulation results show that compared with the traditional constant step conductance increment method, the new MPPT (maximum power point tracking) control method shows faster convergence speed and higher convergence accuracy, and effectively overcomes the limitation of the traditional method in processing optimization speed and accuracy.

With the development of power construction, due to the traction plate passing over the pulley, the pulley wheel diameter does not match the traction plate, resulting in jamming and causing the pulley to fall off, as well as phenomena such as wire jumping and wire jamming during the stringing process. In order to study the problems of jamming and derailment when the traction plate passes over the pulley during tension unwinding, this paper selects two typical traction plates, namely the sheep horn traction plate and the traditional traction plate, and conducts dynamic simulation and comparative analysis of the two typical traction plates passing over the unwinding pulley. The simulation results show that the impact force of the sheep horn traction plate passing through the pulley is smaller than that of the traditional traction plate, verifying that this simulation method is in line with the actual situation. The simulation method summarized for typical traction plates can provide technical foundation for future traction plate structural design.

In response to the challenges posed by high labor costs and low mechanization in the harvesting of safflower filaments in Xinjiang, this study introduces an intelligent detection method leveraging YOLOv8s. A safflower filament dataset was developed and enhanced as the basis for constructing a detection model that incorporates the C2f, SPPF, and Detect modules, and the Loss function. The model was evaluated using recall rate (R), precision (P), and mean average precision (mAP) as metrics. We compared 12 variants of target algorithms based on YOLOv3, YOLOv5, YOLOv6, and YOLOv8. The findings indicate that YOLOv8s achieved a precision of 82.8%, a recall of 78.2%, and mAP of 86.2%. Relative to YOLOv3-tiny and YOLOv5s, YOLOv8s demonstrated higher recall and mAP. Despite its compact size of only 5.96MB, YOLOv8s exhibited superior confidence with no missed or false detections compared to these models. To further affirm the reliability of YOLOv8s, its detection performance on safflower filaments was tested under various conditions, achieving mAP values of 91.8%, 92.8%, 90.3%, 79%, and 92.5% respectively, showcasing its rapid and accurate detection capabilities while maintaining lightness and robustness, potentially serving as a technical reference for the development of intelligent safflower harvesting robots.

Surface damage on railway rails is a critical factor affecting train safety. To address the low efficiency and poor precision in detecting small defects like cracks and rust, a rail damage detection algorithm combining MobileNetv3 and Transformer is proposed. This algorithm embeds spatial coordinate information into channel attention and integrates the CA-Benck module into MobileNetv3 to enhance feature extraction and generalization. Then, Utilizing MobileNetv3, CA-Bneck, and the Transformer encoding module, we crafted a streamlined backbone network dubbed MobileNetV3-CATr. Finally, a BiFPN-Lite module is added to capture more defect features without increasing complexity. Through the YOLO Head, it outputs rail damage information. Experiments on our rail dataset show that our model achieves a mAP of 93.8% at 19.5 FPS, outperforming YOLOv5 by 6.5%, enabling high-precision detection of rail surface damage.

This paper addresses the problem of state estimation for nonlinear discrete systems with correlated noises and one-step randomly delayed measurements. The one-step randomly delayed measurements are modeled by Bernoulli random variables. Utilizing Bayesian filtering theory, an optimal estimation method is derived through constructing an orthogonal transformation matrix to eliminate noise correlation, resulting in a recursively derived Gaussian filtering framework. The weighted integral in the filter is computed by using the sphere volume rule, leading to the specific implementation form of the improved cubature Kalman filter algorithm within the filtering framework. Furthermore, when system noise is independent and real-time measurements are obtained, this framework can degenerate into a conventional Gaussian filtering framework by setting both noise correlation coefficient and probability of one-step random delay to 0, thereby enhancing universality compared to existing results. Simulation results demonstrate that the proposed improved CKF exhibits better estimation accuracy and robustness.

Knowledge Enhanced Pre-trained Language Models have recently improved context-aware representations through external knowledge and linguistic knowledge from grammatical or syntactic analysis. The mismatch between text and knowledge graph embeddings in the feature space cannot be resolved in the fine-tuning phase and input module knowledge augmentation. In this paper, we revisit and advance the development of natural language understanding in Chinese and propose an improving Chinese knowledge Enhancement Models with Unified Respresentation Space. Specifically, knowledge embedded in the knowledge graph triples is effectively injected into it based on a novel pre-training task and knowledge-aware masking strategy. We conducted extensive experiments in seven Chinese nature language process tasks to evaluate the proposed model. The experimental results show that Our Model understands the external knowledge more deeply. We also demonstrate the effectiveness of the proposed method by ablation experiments.

In this paper, we borrow the idea of AdHoc to improve the traditional AdHoc network into a hierarchical network, i.e., each core node only serves as an access point and transmission point, and the real terminal is still the miner's handheld terminal. Firstly, a number of wireless access points (WPs) are randomly arranged in the tunnel, and the WPs accomplish two functions: one is to communicate with users, that is, mobile terminals (MTs, mobileterminals), which are required to have the wireless access and frequency allocation functions of a base station; the other is to complete the data storage and forwarding between them and other WPs, similar to the router function in a computer network. Then, a selforganizing network based on the improved ant colony algorithm protocol communication routing is proposed, which takes the movement speed and load situation of the core nodes as the routing considerations and proposes a routing performance function to measure the merits of communication routing, and the algorithm establishes two communication routes, the primary route and the backup route, simultaneously. Finally, for the special environment of the tunnel, the dynamic routing table is established with the lowest loss between two points by considering channel fading and multipath. Simulation results show that the self-organizing network preference method based on hierarchical model and ant colony algorithm proposed in this paper can improve the success rate of packet forwarding, reduce the average path delay, save the resources of the network, and thus improve the performance of the network.

In wireless sensor networks, numerous nodes deployed in a distributed manner are vulnerable to malicious attacks. To ensure that nodes receive high-level secure data information services and provide secure data query services to legitimate users, this study put forward a high security WSN node bidirectional authentication algorithm. On the basis of establishing an energy model for wireless sensor networks, clustering techniques are adopted to WSN nodes to cut down upon energy consumption. As part of the preliminary authentication process, network initialization, key negotiation, and node trustworthiness calculation are performed on cluster nodes to obtain trusted nodes. By adopting digital authentication methods, bidirectional authentication between trusted nodes has been achieved, which heightens the security level of wireless sensor networks. As repeatedly demonstrated by experimental findings, this method effectively elevates the number of surviving nodes in the network while reducing the authentication delay and energy consumption of WSN nodes.

Although frequency domain analysis theory has been widely used in many signal analysis fields and some practical examples have been reported, the traditional automatic test systems have limited signal analysis ability, thus frequency domain analysis methods have not been widely used in automatic test systems for engineering practice. Aiming at the above problems, we first complete the frequency domain test function requirement analysis in the automatic test system to integrate frequency domain analysis theory into engineering practice, and then determine the signal processing tools commonly used in the analysis frequency domain test process. After that, according to the data characteristics of all kinds of signals in the system, the data structure design is constructed. Finally, the functional principle of the signal processing module is analyzed, and a series of signal processing tools such as algorithm compilation are designed. The signal analysis ability of the system is enhanced, so that the collected data can be processed by the system besides driving the instrument to carry out the test. As a result, the frequency domain characteristics of the tested object are obtained, which provides a basis for fault diagnosis of board-level circuits.

Pneumonia is one of the leading causes of illness and death worldwide. In clinical practice, Chest X-ray imaging is a common method used to diagnose pneumonia. However, traditional pneumonia diagnosis through X-ray analysis requires manual annotation by healthcare professionals which delays diagnosis and treatment. This study aimed to investigate and compare three different deep learning methodologies for classifying pneumonia to detect the disease in patients. These advanced models have the potential to overcome the challenges of reliability and accessibility of diagnostic practices. The methodologies evaluated included a custom convolutional neural network (CNN), a transfer learning approach as well as a fine-tuning strategy based on ResNet152V2. The models were rigorously assessed and compared across various metrics, including testing accuracy, loss, precision, F1 score, and recall. The comparative analysis shows that the fine-tuning strategy outperforms the other methods in terms of operational effectiveness, with the custom CNN being the next most effective, and the transfer learning method ranking last. The study also highlights that false negatives can have more serious consequences than false positives, even without specialized medical knowledge.

This paper presents the design of a rotational varifocal metasurface lens based on the principle of moiré lenses, utilizing square single-crystal silicon pillars. The metalens comprises two cascaded face-to-face metasurfaces, enabling focal length adjustment through the mutual rotation of these two metasurfaces. This design achieves high transmittance and complete 2π phase coverage. And through simulation verification, by rotating one layer of the metasurface from 60°, 72°, 90°, and 120°, the focal length of the designed moire metalens shifts from 5.68mm to 2.39mm along the -z direction on the xz plane, proving that the focus-shifted metalens we designed can be controlled by independently rotating the single-layer metasurface to adjust the focal length within the same focal plane.

It can promptly detect hot-spot in the operation of the generator by real-time monitoring of the operating temperature of the windings in the hydromotor, thereby ensuring the safe operation of the hydromotor. To overcome the limitation of the sensing temperature measurement distance of the down-conversion fluorescence optical fiber temperature sensor does not exceed 30 meters, a method was proposed to use a semiconductor laser with the wavelength of 1550nm which was a low loss in optical fiber as the excitation source to generate upconversion fluorescence for an optical fiber temperature sensor. Upconversion materials ZnF₂: Er³⁺, Yb³⁺ were selected as the fluorescent materials for the sensor. An experimental device was constructed, and the upconversion-fluorescence lifetime and the temperature were tested and calibrated. The results showed that the optical fiber temperature measurement distance of this method can reach to 76m, meeting the demand for the temperature monitoring of the hydromotor windings.

The thermal ablation and the cutting tissue for tumors with laser surgery are very important in nowadays medical surgery. Therefore, the real-time monitoring of the scalpel temperature has become the focus technical research. In this paper, a laser fiber scalpel temperature measurement system based on the intensity peak ratio of upconversion fluorescence excited by 980nm laser was designed, and a type of upconversion fluorescent material was demonstrated and applied to the optical fiber scalpel, so that the accuracy and safety of surgical temperature measurement can be improved. This paper firstly analyzed the principle of upconversion fluorescent temperature measurement. Secondly, the preparation of the fluorescent material was introduced and the CaWO₄: Er³⁺, Yb³⁺ was placed in a quartz tube to make a fiber-optic sensing probe, and the fluorescence intensity peak ratio(FIR) temperature measurement was carried out by warming up (up to 800℃), cooling down, and warming up again to calibrate the relationship between the temperature and the FIR, and the fourth-order expression of the FIR with the temperature was obtained. The fluorescent material was combined with the optical fiber scalpel tip, which was used to improve the accuracy and safety of surgery by changing the power of the laser. Then, the fluorescent material was combined with the scalpel tip, and the laser optical fiber scalpel was heated by increasing the power of the 980nm laser. Finally, the FIR thermometry value was calibrated with the temperature to obtain the laser power-temperature equation, and the temperature accuracy was ensured by multiple measurements to realize the real-time monitoring of the temperature of the optical fiber scalpel in the process of optical fiber laser surgical scalpel (OFLSS).

This paper proposes a kind of three-axis automatic drilling-hole equipment for inner flap beam products. The equipment is mounted on a common platform by using a tooling guide rail system to realize the function of automatic drilling-hole. The equipment adopts lightweight design which has compact structure and simplicity of operation, including tooling guide rail system, three-axis motion mechanism, drilling-hole end actuator, control system and software system. In this paper, the design scheme of the equipment is proposed, and the optimum arguments are determined through the experiment of drilling-hole, which ensures the efficiency and quality of drilling-hole, and realizes the automatic drilling-hole of aircraft beam products.

This article investigates the current development status of agricultural information collection systems and proposes the construction of a smart agriculture system based on a three-layer framework of the Internet of Things. The perception layer collects sensor data of the crop planting environment with RS485 interface, uses short distance wireless communication technology ZigBee for networking, uses wireless interrupt DTU for network layer access, configures DTU to work using HTTP and MQTT protocols, uses Web for data display at the application layer, deploys MQTT servers to complete message subscription and publishing, and elaborates on the key technologies used in the construction of the agricultural system, laying the foundation for the establishment of growth models and digital twins for subsequent data collection in the agricultural system.

In order to study the influence of deck shaking during ship motion on the stability of ship deck foreign matters cleaning robot, based on the random wave model and the ship's motion with the wave, MATLAB software was used to simulate and analyze the random wave, the ship's motion with the wave and the force of the ship deck foreign matters cleaning robot, and satisfactory simulation results were obtained. In this paper, the stability of ship deck foreign body cleaning robot is analyzed, and a stability analysis method is proposed, which provides a theoretical basis for the stability analysis of ship robot.

Video target tracking has a wide range of application value in the field of automatic driving, UAV target tracking, security monitoring, etc. How to maintain stable tracking of the target among video data frames is the focus of the research. A robust tracking algorithm that effectively solves the target drift problem is proposed for the problem of target loss due to image perturbation, scale change, target occlusion and other disturbances when the KCF algorithm is used for video target tracking. The algorithm is based on the KCF algorithm framework, which proposes a multi-scale sampling strategy and designs a multiple classifier screening algorithm to ensure the accuracy of the target template. Through experimental verification, the algorithm can effectively solve the drift problem in the tracking process and realize the continuous accurate tracking of the target. The algorithm provides a real-world application reference for engineering applications of real-time video data processing.

While Radio Frequency Identification (RFID) has become an indispensable technology, there are many questions about the security it brings. RFID technology is represented by the ISO 18000-X technical standards. These define the essential physical layer and communication protocol data required for the exchange between interrogators and transponders. These standards are organised by frequency. In this article, we present a system that automatically detects, identifies; and verifies the presence/absence of students during an exam without using call lists. We will begin with an introduction to the technology, the architecture of an RFID system, its performance; and specific characteristics, and end with a presentation of our system.

With the development of robotics and self-driving vehicles, simultaneous localization and mapping technology has become critical. Visual SLAM utilizes the camera as the primary sensor and shows significant advantages in texture-rich static environments. However, conventional vision SLAM systems do not perform well in dynamic environments, such as changes in lighting, loss of camera positioning, or movement of feature points due to object motion. This study proposes a new SLAM system for dynamic environments, YOEC-DSLAM. YOEC-DSLAM combines a lightweight semantic segmentation network and a multi-view geometry approach to effectively identify and filter dynamic feature points, thereby significantly improving the accuracy and robustness of map construction. These improvements improve the real-time and accuracy of the system across the board. Experiments show that YOEC-DSLAM exhibits unique advantages in processing highly dynamic scenes.

Unsteady flow fields are widely present in nature and engineering, but their dynamic characteristics and complexity make the process of solving flow fields a complex and time-consuming task, generally requiring the use of Computational Fluid Dynamics (CFD) methods. However, this method has problems such as large pre-processing workload, long computation time, slow convergence speed, and limited prediction accuracy. To solve these problems, this paper combines the hard manner boundary conditions physical information network with the deep residual networks (ResNets) to propose a new solution method for reconstruction and prediction of flow fields. Taking the advection equation and two-dimensional cylinder wake problem as examples, the solution results are compared with CFD and (Physics-Informed Neural Networks) PINN. The experiment proves that this method can solve unsteady flow fields with high accuracy, and the accuracy is improved by about 95% compared to PINN. It also has good stability and robustness, providing a new approach for solving fluid mechanics problems.

Microstructure evolution and hardening mechanism of 18Cr2Ni4WA steel subjected to combined high-temperature carburizing, two high-temperature tempering, carbonitriding, quenching, low-temperature treatment, and low-temperature tempering heat treatment were investigated in this study. The effect of surface precipitation phases and retained austenite on surface hardness after different heat treatments was studied throughout the process. The results indicated that the combined treatment converted the lath martensite (LM) surface layer into acicular martensite (AM), carbide and residual austenite (RA). Furthermore, subzero treatment led to the improvement of mechanical properties through the reduction of residual austenite and the increase of carbide. The deformation of the martensite lattice and the compressive stress within the retained austenite are reduced through low-temperature tempering, which led to a decrease in austenite stability, facilitating its transformation into martensite. This study deepens the current understanding of the hardening mechanism of 18Cr2Ni4WA steel during carburizing carbonitriding composite heat treatment, providing a theoretical basis and experimental basis for expanding its application fields.

In order to accurately predict the vibration characteristics of gears, a refined dynamic model for gears is proposed in this study. A novel bearing collision model is established, incorporating nonlinear oil film forces to better simulate bearing collisions during transmission. Additionally, a new gear fractal tooth clearance model is established based on fractal theory and combined with tooth surface roughness. The gear dynamic model is solved using a closed-loop algorithm by combining viscous damping factor and time-varying pressure angle. The responses of different models are compared by the bifurcation diagram, phase portraits, and Poincaré mapping. The result shows that the proposed new bearing model accurately describes bearing collision phenomena compared to traditional bearing models. As the restitution coefficient increases, the system response gradually transitions from periodic to chaotic motion. The stability of the system response decreases with the increase of tooth surface roughness and fractal dimension. Compared with fractal dimension, tooth surface roughness has a more obvious effect on the dynamic response of the system.

This study analyzed the performance of FTO/TiO2/FAPbI3/CZTS solar cells using SCAPS-1D simulations. Various parameters, including thickness, acceptor density, and defect density for the FAPbI₃ and CZTS absorber layers, series resistance, shunt resistance, and operating temperature were optimized to enhance device efficiency. The structure consisted of a FTO front contact, TiO₂ as the electron transport layer (ETL), FAPbI₃ as the perovskite absorber layer, and CZTS as the secondary absorber layer. The optimized parameters for achieving the highest efficiency included an FAPbI₃ thickness of 0.8μm, a CZTS thickness of 2μm, and an acceptor density of 10¹⁶cm^-3 for FAPbI₃ and 10¹⁸cm^-3 for CZTS. A defect density of 10¹⁴cm^-3 for both absorber layers and an operating temperature of 300K. Increasing series resistance leads to a decrease in fill factor and efficiency. In contrast, higher shunt resistance enhances the fill factor and efficiency. Under these conditions, the solar cell achieved an open-circuit voltage (V_OC) of 1.1148V, a short-circuit current density (J_SC), an overall efficiency (Eta) of 26.17%, and a fill factor (FF) of 87%. These findings contribute significantly to optimizing perovskite/CZTS-based solar cells for improved performance.