PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
This PDF file contains the front matter associated with SPIE Proceedings Volume 13161, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Fiber optic sensors due to their advantages such as light weight, small size, high temperature resistance, corrosion resistance, and electromagnetic interference resistance, can largely overcome the influence of environmental factors, accurately monitor various parameters of aircraft, and timely judge and prevent accidents. Therefore, they are suitable for parameter monitoring of unmanned aerial vehicles. In the working state of unmanned aerial vehicles, optical fibers are easily affected by axial forces, which in turn affects the accuracy and stability of fiber optic sensors in monitoring aircraft parameters. This study applies genetic algorithm to establish a mathematical model for the axial force on optical fibers. The algorithm is implemented through steps such as determining decision variables and constraints, establishing an optimization model, encoding, decoding, selection, crossover, mutation, and population iteration. Finally, by changing different parameter values for simulation analysis, the optimal approximate solution of the center wavelength value when the fiber optic is subjected to axial force was obtained. This effectively adjusts the fiber optic sensing system, improves the real-time, accuracy, and stability of unmanned aerial vehicle parameter monitoring, and has good application value and promotion significance.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
A new structure of photonic crystal fiber with high birefringence and high nonlinearity is proposed. The birefringence, effective mode area and nonlinear coefficient characteristics are studied by finite element method (FEM). Numerical studies show that when the cladding hole spacing of fiber is 1.0μm, the birefringence is 2.06×10-2 and the high nonlinear coefficient of 69 W-1km-1 can be obtained at the wavelength of 1.55μm. This kind of fiber is easy to fabricate because of its simple structure and strong regularity, which has a wide application prospect in polarization control.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
An all-fiber polarization maintaining laser system of wavelength tunable from 1.76 to 1.84 μm based on the Raman-induced soliton self-frequency shift in an erbium-doped amplifier is demonstrated. The system only includes oscillator and two-stage amplifiers which are built entirely based on commercial silica fibers and devices. The ultra-short pulse achieved by oscillator and first amplifier stage by the method of nonlinear compression is delivered into second amplifier stage which is designed as both Raman shifter and amplifier. Within the pump power range of 16-25 W, tunable Raman soliton can be obtained with the average power gradually increases from ~180 mW to ~250 mW as the center wavelength of the soliton shifts towards the longer wavelength direction within the range of 1.76 to 1.84 μm. The system delivers Raman soliton from an entirely fiberized, fusion spliced system without any free-space optics, which can provide robust and stable soliton generation. To our knowledge, this tunable soliton source is the highest output power demonstrated within 1.7-1.8 μm wavelength range from an all-fiber polarization maintaining laser. Our experiment provides a feasible femtosecond source for multiphoton microscopy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The displacement measurement system consists of an edge finder, a displacement drive system, and a grating. The system structure is complex. Magnetic levitation ruler can achieve both driving and detection functions simultaneously. The subdivision of the output signal of the magnetic levitation ruler can improve the detection accuracy of the magnetic levitation ruler. According to the characteristics of the magnetic levitation ruler, under the condition of constant magnetic induction intensity and horizontal control coil current, the displacement of the mover core is proportional to the square of the movement time. This article proposes a dynamic and nonlinear grid subdivision method based on the time-space conversion method. This article uses TDC-GP22 to set the length of the time grid, which can achieve 1000 times subdivision of the output signal of the magnetic levitation ruler. The subdivision of the output signal of the magnetic levitation ruler improves the measurement accuracy of the magnetic levitation ruler.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
A reservoir computing system (RC) based on a vertical cavity surface emitting semiconductor laser (VCSEL) subject to arbitrary-polarization optical feedback is proposed, and its performance for predicting Santa-Fee time series is numerically analyzed. Taking advantage of the fact that the light emitted by VCSEL has two polarization components, a tunable polarization wave plate is placed in the feedback loop to adjust the polarization angle of the feedback light. The influences of system typical parameters on the prediction error of such a RC system have been analyzed in detail. Through optimizing the parameters of feedback strength, injection coefficient and polarization angle, the prediction error of 3% can be realized for the reservoir computing system.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Currently, the Kalman filter can only be applied to linear systems. The extended Kalman filter and unscented Kalman filter developed from the Kalman filter are widely used in solving nonlinear system filtering after strapdown seeker decoupling. With the development of precision guided weapons, the full strapdown seeker has been widely used in many guided weapons. This paper analyzes the working principle of the strapdown seeker, discusses the existing problems of the strapdown seeker, and further introduces the scheme of line of sight Angle rate extraction, and introduces the line of sight Angle rate estimation model through the relationship between coordinate conversion and angular rate decoupling. Through extended Kalman filter, unscented Kalman filter, The filtering principle is analyzed and the eye Angle rate is estimated and simulated by Matlab. The simulation results show that the accuracy of unscented Kalman filter is higher than that of extended Kalman filter, and the error is smaller. Compared with EKF, UKF is more suitable for filtering estimation of nonlinear system of strapdown seeker. Finally, by analyzing most of the existing filtering algorithms, the prospect of filtering algorithm research is put forward.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
According to the combustion spectra of hydrocarbons, three infrared flame sensors (RD-913FB1, RD-913FB2, RD- 913FB3) based on pyroelectric effect and ultraviolet flame sensor (R2868) based on photoelectric effect are used to develop a flame detector which can quickly detect the surrounding flame by means of its powerful calculation and control capability. By adopting the composite detection technology of four wave band sensors, the detection range of the flame radiation spectrum of the system is enlarged; amplifying and shaping the signal of the sensor by calculating and designing an amplifying filter circuit; The intelligent flame detection algorithm is designed according to the site environment, which can effectively judge whether the surrounding environment produces a fire or not, and quickly carry out acousto-optic alarm. Compared with traditional temperature and smoke flame detectors, the accuracy is also improved. The 0.4 m standard flame is used for the ignition experiment, and the external environment is changed to use automobile headlights, sunlight and artificial light sources as interference items to carry out anti-interference detection on the flame detector. The experimental results show that the flame detector has the characteristics of fast response speed, long detection distance and strong anti-jamming ability, which fully meets the design requirements.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The geological survey along the large railway in China is difficult, the boundaries of many regional geological maps are unclear, and the geological mapping is very different. The use of hyperspectral remote sensing can greatly improve the survey accuracy in complex mountainous areas. In the identification of lithology and structure, hyperspectral remote sensing has certain advantages compared with multi-spectral satellite images. In this paper, the hyperspectral data of Gaofen-5 satellite is interpreted by a series of technical methods, such as data preprocessing, radiometric calibration, atmospheric correction, orthographic correction, image information enhancement, radiation homogenization processing, etc. According to different geological conditions, lithologic structure information by hyperspectral information enhancement technology in different regions and layers. Different geological units are effectively divided, the distribution characteristics of geological structure space are identified, and some stratigraphic boundaries of the original geological map are adjusted. This paper solves the problems that the survey personnel of long and large tunnels cannot reach, and the conventional survey methods cannot verify the stratigraphic boundaries and lithological characteristics. It also provides an important auxiliary reference value for the identification of surrounding rock of hyperspectral remote sensing.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In the context of photonic crystal fibers used in the field of communication. In this paper, a new design of the photonic crystal fibers is designed and proposed. The analysis of the band-gap of gallium lanthanum suphide glass hollow core photonic crystal fiber is conducted using the plane wave expanding method. Furthermore, the properties of confinement loss and velocity dispersion of the waveguide groups of this kind of fiber are investigated using multiple methods. It is discovered that a band gap of 0.15μm in the mid-infrared band is present in this fiber, and the diameter of the cladding air hole has a significant impact on the confinement loss and the waveguide group velocity dispersion. Last but not the least, we got the confinement loss measures 0.0169dB/m at a wavelength of 2.7μm. This PCF is suitable for research in the mid-infrared wavelength range. This design has great application significance in the fields of fiber optics and optical communication of infrared band.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The real-time measurement of low-level wind fields has very important practical significance and application value in fields such as climate, agriculture, and engineering applications. Especially, the accurate acquisition of wind field information within a height of 200 meters has become an urgent problem to be solved. The developed paraxial Doppler coherent wind radar system uses a continuous laser light source with a wavelength of 1550 nm that is safe for the human eye, and adopts an all fiber coherent detection method. By separating the transmission and reception of laser signals, the design further reduces the backward reflection of the fiber end surface, thereby enhancing the ability to identify Doppler shifts in the echo signal. Compared with the wind speed simulated by the runner device at a fixed position, the linear correlation coefficients between the simulated wind speed and the measured wind speed along the fitting curve are both greater than 0.999, verifying that the system can effectively achieve the measurement of wind speed. The system can detect designated areas of the atmospheric wind field at different distances, and can also quickly and stably detect wind speed signals. The relevant research can be used to retrieve the actual atmospheric low-level wind field for real-time monitoring.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
This study investigate the integration of discrete modulated coherent state (DMCS) continuous variable quantum key distribution (CV-QKD) with wavelength division multiplexing (WDM) in 5G fronthaul optical networks. Firstly, an architecture of quantum-classical co-propagation in C-band DWDM-based 5G fornthaul network is constructed and the wavelength assignment scheme for classical/quantum channels is designed. By assigning the quantum channels in band of low Raman scattering coefficient and setting the quantum-classical channel space to be 200 GHz, the proposed scheme can effectively reduce the spontaneous Raman scattering (SpRS) and four-wave-mixing (FWM) noise. Then, the secret key rate (SKR) of DMCS protocol and the excess noise induced by SpRS are analyzed theoretically to furtherly character the co-propagation architecture. Lastly, numerical simulation is carried out to evaluate the impact of SpRS noise on the CV-QKD performance. The simulation results show that by employing the proposed wavelength assignment scheme, the co-propagation system can support positive SKR at a maximum distance of approximately 41 km for 8 classical channels use and 30 km for 16 classical channels use, respectively. Our work can provide a reference for practical deployment of quantum secured 5G fronthaul optical networks.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In recent years, with the rapid development of satellite communication system technology, various countries and organizations are paying more and more attention to the field of low earth orbit (LEO) satellite communication, such as Star Link products. This paper presents a scheme of TT&C equipment based on TT-1 satellite to meet the data transmission requirements of explosively enhanced LEO satellites. By calculating the orbit ephemeris and receiving the satellite broadcast signal to implement terminal manage and compensate for the Doppler frequency offset, simplifying and optimizing the protocol to achieve fast access to the network and beam handover, using anti-irradiation technology to prevent single event upset, the key technology of TT&C terminal is researched and realized. Compared with the traditional TT&C method, the TT&C technology prototype proposed in this paper has the characteristics of low construction cost, continuous data transmission, and long service time, which can well serve the booming LEO satellite industry in the future.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Amidst the intricate battleground where attackers and defenders are perpetually locked in a sophisticated digital cat-and-mouse game, the task of discerning malware within an adversarial crafted environment manifests escalating complexity. The current research delineates the utilization of an enhanced reinforcement learning technique, orchestrated to engender adversarial malware specimens, thereby strategically navigating through machine learning detectors. This endeavor not only bolsters the robustness of malware identification systems but also adeptly navigates the perpetually evolving machinations of malware creators. Within this research, an environmental model is meticulously constructed, emulating detection engines and feature extractors, with malware samples assimilated as input. By integrating an autonomously generated reward function, we ascertain the model’s agility and the concomitant generation of manifold adversarial malevolent samples. The empirical evaluations underscore that, in contrast to conventional machine learning approaches, our methodology exudes superior flexibility and efficacy, furnishing a more formidable challenge to malware detection mechanisms.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
A multi parameter monitoring system based on STM32 microcontroller was designed to address the security risks in computer laboratories, achieving dynamic monitoring and real-time alarm of the laboratory environment to improve the safety and management level of the laboratory. Parameters such as humidity and gas concentration were used to monitor the laboratory environment. The STM32 microcontroller serves as the control core of the system, controlling and collecting data from various sensors through analog input and output ports. At the same time, the system adopts network communication technology to transmit data to the upper computer for processing, achieving abnormal alarm function. The experiment shows that the system can accurately monitor the environmental parameters of the laboratory and transmit data to the host for processing through network communication technology, achieving abnormal alarm function. During the real-time monitoring process, the system can timely detect and alert laboratory safety hazards, effectively improving the safety and management level of the laboratory. The system has achieved real-time monitoring and abnormal alarm of the laboratory environment, improving the safety management level of the laboratory. This system is practical and feasible, and can provide reference for laboratory safety management.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Federated learning can utilize its distributed structure to protect data privacy security of clients and improve efficiency of machine learning. However, its distributed framework also make itself be susceptible to sybil attacks. While previous research has already proposed defense methods to address this issue, they often fail to guarantee effective performance in a dynamic federated learning system, where some clients dynamically join in and out. To tackle this problem, our paper introduces a novel defense method specifically designed to mitigate sybil attacks in dynamic federated learning scenario. Our proposed method consists of three mechanisms: similarity mechanism, validation mechanism, and reputation mechanism. These mechanisms can address the problem of missing information and effectively resist sybil attacks in dynamic federated learning. We evaluate the performance of our method on the MNIST and KDDCup datasets and demonstrate its advanced ability in defending against sybil attacks in dynamic federated learning compared to existing methods.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The interrupt response speed can reflect the CPU's processing speed for external events, and is a very important indicator for measuring CPU performance. In the field of real-time applications, interrupt response speed is an important manifestation of real-time performance and an important aspect of optimizing CPU implementation. When responding to interrupts, especially in CPU implementations that allow interrupt nesting, it usually involves saving the interrupt context. Only after saving the interrupt context can the CPU execute instructions related to interrupts in the interrupt handling function. This paper presents a hardware implementation scheme for backup registers. One application method is to quickly save interrupt context information to the backup registers when responding to an interrupt, and then save the contents of the backup registers to the stack parallel in the background when executing interrupt processing function instructions to improve interrupt response speed. Another application method is to restore the interrupt context from the backup register directly, or to obtain the current interrupt context information from the current interrupt stack position in the background in parallel during the execution of the interrupt processing function instruction to the backup registers. When the current interrupt returns, the interrupt context can be directly restored from the backup registers to improve the speed of interrupt return. Hardware stack identification can be used to indicate whether interrupt nesting can be performed or not. Whether interrupt context can be restored from the backup registers or not can be checked by the effective hardware identification for backup registers information. The process of parallel reading interrupt contexts from stack to backup registers in the background can be controlled by recording the stack position corresponding to each interrupt context in hardware.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
This paper proposes an improved genetic algorithm for dynamic resource management, taking into account network delay and energy consumption. The algorithm utilizes CloudSim and CloudAnalyst tools to analyze qualitative and quantitative its performance. The experimental results demonstrate that the algorithm reduces response time for user requests and improves Quality of Service (QoS) while consuming the same amount of power. Additionally, It also leads to lower power consumption for the same response time. This research finding is significant for enhancing the performance and efficiency of cloud computing platforms. The work holds practical value as it offers an effective solution for resource management in cloud computing environments. This study proposes an improved genetic algorithm that optimizes resource allocation, reduces network delay, improves energy efficiency, and enhances user experience in cloud computing technology. The algorithm is innovative and practical in solving dynamic resource management problems, providing valuable references for related research fields.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
With the pervasive use of the Internet, recommendation systems have gained increasing importance in people's daily lives. Among the crucial tasks in recommendation systems, click-through rate prediction stands out as it directly influences their effectiveness. Recent studies have revealed that incorporating user behavior sequences can substantially enhance the accuracy of click-through rate prediction models. However, existing models overlook user preferences for textual and visual information, which hampers the acquisition of a comprehensive representation of user interests. Consequently, this limitation results in suboptimal model accuracy. In this paper, we propose a unified framework for modeling multi-modal user behavior sequences. Our framework leverages a unified cross-modal pre-trained model for feature extraction and employs a multi-modal similarity-enhanced attention mechanism to capture users' preferences across various modalities. We conduct extensive experiments on large-scale real-world datasets to validate the effectiveness of our approach. Compared to other state-of-the-art click-through rate estimation algorithms, our model achieves an approximate 1.22% improvement in AUC, thereby significantly enhancing the accuracy of the recommendation model.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Aiming at the problem of time loss caused by weight distribution imbalance in ALNS module of MAPF-LNS2 algorithm, ALNS+ module is proposed by introducing improvement rate statistics, time window, improvement rate trend judgment function and other mechanisms. Compared with the ALNS module, the ALNS+ module considers the recent improvement rate trend of the neighborhood search strategy, and switches to other neighborhood search strategies when the trend decreases, so as to repair the excessive weight allocation in time and reduce the time loss. The experimental results show that the ALNS+ module achieves a significant improvement in running time, and the maximum reduction is 65.1%. However, in scenarios such as denser agents, the execution success rate of the improved module has a significant downward trend. The PNS module was proposed to speed up the neighborhood repair process by introducing a parallelization method to parallel the neighborhood search process on multiple processor cores. The comparative experimental results show that the PNS module has a significant improvement in the success rate, with a maximum increase of 40%. By integrating the ALNS+ and PNS modules into the MAPF-LNS2 algorithm framework, the MAPF-LNS2* algorithm was proposed. The simulation results show that the MAPF-LNS2* algorithm effectively solves the problem of time loss in the MAPF-LNS2 algorithm, and effectively reduces the failure rate in the agent dense scene.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Conventional deep learning based object detection methods demand substantial annotated data for training, incurring considerable time and labor costs. Conversely, few-shot object detection necessitates only limited data from novel categories, emerging as a prominent research focus. This study proposes the Attention Contrastive Network (ACNet) to address few-shot object detection challenges. ACNet incorporates an attention mechanism architecture, extracting attention values and keys from image features in both support and query sets. It compares key attention across the sets and weights query set features with attention to augment local features. Additionally, multi-scale pooling layers enhance the network's capability to identify objects across varying scales. The introduction of an attract-repel mechanism in the loss function significantly amplifies inter-class differences, thereby improving classification accuracy. ACNet's efficacy is experimentally affirmed on the PASCAL VOC and COCO datasets, yielding commendable results in few-shot detection tasks.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The achievements in modern unmanned systems technology are noteworthy, with Unmanned Ground Vehicles (UGVs) exemplifying exceptional payload capacity and endurance. The collaborative operation of UGV clusters has emerged as a pivotal direction for constructing future intelligent transportation systems and driving the development of smart cities. Existing algorithms predominantly focus on single-agent systems, leaving a research gap in the exploration of multi-agent systems. To address this void, this paper concentrates on resolving the UGV cluster search problem in closed road sections. The approach involves the utilization of multi-agent reinforcement learning algorithms to handle task scheduling and collision avoidance, enabling UGV clusters to complete extensive area search tasks with minimal data transmission efficiently. This methodology ensures the decentralized and efficient operation of UGV clusters, showcasing robust search capabilities in closed road sections. The practical applications of this approach underscore its substantial potential.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
A multi-scale traffic scene object detection algorithm based on Light-YOLOX has been proposed with the aim of achieving a substantial enhancement in accuracy while maintaining a lightweight design. To accomplish this, the algorithm has undergone further refinement by introducing a pyramidal attention segmentation module in the backbone network and pooling structure. This enhancement has significantly improved the algorithm’s capacity to extract contextual information at various scales, leading to more precise and comprehensive object detection in traffic scenes. Moreover, the development of the Py-FPN feature fusion structure, achieved through the integration of multi-scale pyramidal convolution, has enabled the complete fusion of output feature layers from the backbone network. This integration has further heightened the algorithm’s detection effectiveness, ensuring a more robust and accurate detection of objects within traffic scenes. Through experimental evaluations on both the KITTI dataset and Cityscapes dataset, it has been demonstrated that the proposed multi-scale traffic scene object detection method delivers a considerable improvement in accuracy and comprehensive performance, despite having a minimal number of parameters and operations. These findings underscore the algorithm’ s potential applicability and effectiveness in the practical field of traffic scene object detection, illustrating its suitability for real-world deployment and utilization.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Object detection and image recognition are considered to be the most critical and challenging aspects of computer vision. The rapid advancement of driverless technology and assisted driving systems further emphasizes the importance of traffic sign detection and recognition. Detecting small objects against complex backgrounds in practical assisted driving tasks remains a formidable challenge. Many existing target detection methods struggle to strike a balance between performance and parameters. In response, we propose a novel traffic sign detection model based on YOLOv8s—A Traffic Sign Detection Network incorporating Normalization-based Attention and Lightweight Convolution (YOLOv8- NL). The objective is to optimize the performance and address the challenge of balancing model parameters in traffic sign recognition within the transportation domain. Firstly, Normalization-based Attention (NAM) is introduced to address issues related to inconsistent input dimensions, uneven target distribution, and complex background of optimization. Secondly, lightweight Global Sparse Convolution (GSConv) is integrated to reduce the model parameters and enhance model generalization performance. Finally, to further enhance detection accuracy, the Wise-IOUv3 loss function is introduced to tackle difficulties associated with low-quality labeling of data. The effectiveness of the novel YOLOv8 model, which integrates Normalization-based Attention and Lightweight Convolution, is demonstrated through experiments conducted on two publicly available datasets. These experiments showcase the model's ability to significantly reduce parameters without compromising its guaranteed performance. It is noteworthy to note that significant improvements were observed in the CCTSDB2021 dataset, with a reduction of 1.2 million model parameters, resulting in 1% increase in mean average precision (mAP) compared to the YOLOv8s model. Furthermore, the mAP for the three-class target detection task on the TT00K dataset reached an impressive 94.2%.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Aiming at the problem of how to carry out efficient distribution of electric power materials, the optimization model of distribution route for electric power materials considering time window is designed. The model takes the minimum distance travelled by vehicles as the objective, and at the same time considers the constraints such as vehicle load and time window of each demand node, so as to make the model more in line with the demand of electric power material distribution. In order to improve the solving ability of the optimization model, the idea of Genetic Algorithm (GA) and Simulated Annealing Algorithm (SA) are introduced into the Ant Colony Algorithm (ACO) to form the Hybrid ACO algorithm. The feasibility of the algorithm is verified by choosing the international common examples. The results show that the travelling distance of the optimal route obtained by the hybrid ACO algorithm is significantly shorter than that of the optimal route obtained by the GA algorithm. Therefore, the proposed algorithm can provide theoretical references for the decision-making of electric power material distribution.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Hierarchical Federated Edge Learning (HFEL) is a promising and efficient framework , providing privacy preservation, which aims to address the issue of limited resources and network congestion by utilizing the available resources in the edge network. This paper proposes a new scheme, HFEL-Q, based on HFEL to address the high energy consumption problem of FL training, as well as the inherent communication and user heterogeneity problems. To improve training performance, a utility function is designed based on users' learning quality and training time to efficiently select the user group with the highest utility. Additionally, a frequency determination method is employed to optimize idle time and reduce energy consumption during training. Finally, the performance of HFEL-Q is evaluated on two real datasets to demonstrate its superiority over state-of-the-art baselines in terms of training rate, accuracy, and energy savings.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The Xizang script named entity relationship extraction is the foundation and premise of information processing such as machine translation, knowledge graph and network public opinion analysis, the traditional Xizang script named entity relationship extraction based on deep learning may often ignore Category Keywords. This paper applies the Xizang script entity relationship extraction method of graph convolutional neural network (GCN). On the basis of the original word vector, category keyword characteristics are obtained through a keyword acquisition algorithm, and a segmented maximum pooling strategy is used to reduce the loss of information in traditional maximum pooling strategies, the maximum pooling strategy for the pooling process is to select the characteristic with the highest score from a series of characteristic values obtained from each filter in the convolutional layer as the reserved value of the pooling layer. Experiments have shown that the method in this paper significantly enhances the results of extracting Xizang script entity relationships.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
With the widespread use of smartphones and embedded devices, data security issues have become increasingly prominent. Among them, the file system, as a key part of storing and managing data, pl ays a crucial role in its stability and security. YAFFS2 file system has become one of the most commonly used file systems in these devices due to its excellent performance and stability. However, data loss can occur due to various reasons such as accidental deletion or hardware failures. Therefore, researching and developing an effective data recovery technique for YAFFS2 file system is of great significance to ensure the safety of user data. Firstly, we conducted an in-depth analysis of the structure and characteristics of the YAFFS2 file system. YAFFS2 is a file system specifically designed for NAND Flash, which adopts a log structure to effectively handle random writes to NAND Flash. In addition, YAFFS2 also has self -repair and garbage collection functions, further improving its stability and reliability. Next, we introduced common data recovery methods. These methods mainly include file header-based recovery, content-based recovery, and metadata-based recovery. However, when dealing with complex data recovery tasks, these methods often suffer from low efficiency or poor recovery results. To address these issues, we proposed a multi-version data recovery algorithm based on hash chaining and a time series construction method based on timestamps. Both methods are designed based on the characteristics of the YAFFS2 file system. The multi-version data recovery algorithm based on hash chaining can effectively handle multiple versions of files, improving the accuracy of data recovery. On the other hand, the time series construction method based on timestamps can effectively handle the continuity of data, improving the efficiency of data recovery.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
The MapReduce task scheduling mechanism and the HDFS storage strategy in Hadoop are lower in the performance of complex heterogeneous application scenarios, and bave become the botteneck of mass data processing. The addresses based on Hadoop platform design patent image retrieval system which is a concrete, in default of MapReduce task scheduling mechanism and HDFS storage strategy in practical application scene performance is poor, respectively, are proposed optimization measures, improve the performance of the system. The main work is as follows: (1) The design principle and execution flow of Hadoop framework, HDFS and MapReduce are deeply analyzed. (2) Aiming at the problem of the actual use of the Hadoop platform in the actual use of the image patent retrieval system, the improvement of the LASE task scheduling strategy and the HIFI storage strategy are improved respectively. (3) The optimization strategy of the image retrieval for the Hadoop platform is verified by experiments. The experimental results show that the optimized system can improve the performance of the Hadoop platform, and reduce the response time of the users’ request.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Wireless traffic prediction in wireless communication networks holds significant application value for optimizing resource allocation. However, precise prediction faces numerous challenges due to traffic complexity, diversity, and dynamics. Complex spatiotemporal dependencies exist between regional base stations, requiring accurate capture. Traditional centralized methods necessitate centralized storage and processing of substantial user data, posing high privacy risks and potential leaks. This paper proposes a wireless traffic prediction model combining federated learning and differential privacy to capture fused spatiotemporal features of base stations and enhance predictive performance. Concurrently, existing federated learning systems have been proven to harbor potential threats during the training phase, endangering data privacy. This paper tackles privacy threats by adaptively pruning to safeguard data privacy, achieving privacy protection. Experimental results demonstrate that this model outperforms mainstream algorithms in prediction accuracy and privacy protection, showcasing broad application prospects.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In research on knowledge graph representation learning, most of the existing methods use shallow linear models when extracting explicit features, while deep nonlinear models are preferred when extracting implicit features. Although the shallow linear model can learn explicit features, the effect of implicit feature extraction is limited, while the deep nonlinear model can learn implicit features, but it is easy to lead to problems such as too many parameters, overfitting, and loss of explicit features. Aiming at the above problems, this paper proposes a new feature extraction frame-work--- JointMC, which highlights entity-related features through semantic spatial mapping and combines one- and two-dimensional convolutional networks to extract explicit and implicit features in the knowledge graph. JointMC adopts a semantic spatial mapping model to learn the semantic information of the entity, filters irrelevant features, and highlights the features that are closely related to the entity. It then combines the semantic spatial representation with 1D and 2D convolutional networks to extract explicit and implicit features in the knowledge graph. Spatial mapping is combined with 1D and 2D convolutional networks to extract implicit and explicit features. Experimental com-parisons with several models confirm the good performance of JointMC in the link prediction task.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In cognitive electronic warfare, the emergence of multifunctional radars with multiple tasks and working modes of search, confirmation, tracking, identification, and guidance has brought great difficulties to defense and jamming. Especially for the jammer, it is very likely that the radar party will fail to fail and expose its own target if it does not know the intention of the radar party. In order to solve the problems of radar behavior intent identification and threat level analysis of the jammer, we propose a radar threat level recognition method based on the improved support vector machine method of the firefly algorithm, which is of great significance to solve the problems of the exposed target that the jammer may cause by the jammer to start the machine, the difficulty in implementing efficient, real-time and effective jamming, and the threat to the security of the protected area. In this paper, the radar threat level analysis problem is modeled as a classification problem, and for the specific problem that the radar information intercepted by the jammer is extremely limited, considering the good accuracy of the support vector machine method in solving the small sample problem, the support vector machine is selected as the classification model and optimized. Finally, the simulation data experiment is given, and the experimental results show that the support vector machine method optimized by the Firefly algorithm shows good accuracy in the radar threat level analysis problem, that is, the proposed method has good accuracy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In the field of autonomous driving, 3D object detection plays a crucial role as a key perception module. Radar-vision fusion based object detection refers to the technology of combining radar data and visual data to detect and recognize targets. However, in practical situations, visual data may encounter a series of problems during the data collection process, such as limited field of view, lighting variations, motion blur, etc. To address these issues, in addition to the commonly used techniques for onboard cameras, such as dynamic exposure control and low-light enhancement, this paper proposes a novel radar-vision fusion based object detection framework based on CenterFusion. The framework focuses on the case of dealing with abnormal visual data, and aims to achieve more reliable target detection under various complex environmental conditions by fully exploiting the complementary nature of radar and visual data, and introducing a point cloud feature extraction module and a modal attention mechanism. Finally, comparative experiments are conducted on the nuScenes-mini under different conditions, and the experimental results show that the method proposed in this paper can completely replace CenterFusion in various situations, demonstrating excellent performance.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
At present, the technology of the Internet of Things is not yet perfect. Due to the limited resources of nodes in the network, it is vulnerable to attacks and various possible attack methods, which pose great challenges to the information security and privacy of the Internet of Things. How to quickly identify intrusion behaviors in the Internet of Things environment is an urgent problem that needs to be solved in current network security. Based on the working characteristics of Recurrent Neural Network (RNN) and Long Short Term Memory (LSTM) networks, this article optimizes LSTM to obtain GRU (Gate Recurrent Unit) network, and compares it with traditional logistic regression (Softmax) classification methods. The research results of this article show that using dropout to train neural networks can weaken the interaction between neurons and effectively avoid overfitting. The accuracy of GRU-LSTM and GRU-Softmax in the IoT-23 (Internet of Things) database is 96% and 76%, respectively. This article proposes an IoT data intrusion detection method based on the GRU-LSTM algorithm, which has more advantages.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
Aiming at the key problems of farmland irrigation and achieving the purpose of low cost, high energy saving and intelligent irrigation, this paper designs and researches an intelligent irrigation system for farmland based on Internet of Things (IoT).This system makes use of sensors and information technology in the field of water irrigation in the era of Internet of Things (IoT) to improve the automation level of agricultural irrigation, and it is a new embodiment of intelligent water conservation. Through intelligent monitoring, soil moisture and irrigation volume can be precisely controlled to improve the efficiency of water use, thus achieving the purpose of water conservation and playing an active role in alleviating water shortage and related environmental problems. An intelligent irrigation system needs to design the hardware and software of the system. The hardware design of the system mainly includes the design of the collection node module, gateway node module, control module and wireless transmission module. The aim is to achieve real-time data collection and stability of data transmission; in the system software part, the workflow of each node of the wireless communication network is designed to achieve real-time collection of farmland data and precise irrigation of crops.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
In order to improve the processing quality of traditional Chinese medicine rhubarb and control the processing process of rhubarb in real time, a programmable logic control system with Ethernet and Internet of things remote communication function was designed, and the corresponding programming software and configuration software were developed, which had important engineering application value for the transformation and upgrading of traditional processing equipment of rhubarb. In this paper, based on 5G Narrowband Internet of Things (NBiot) technology and industrial Ethernet technology, a programmable logic control system with remote wireless communication function is designed, and the supporting programming software platform is developed to connect to the cloud wireless network. The research results were tested in Li County Rhuai Technology Co., LTD. The rhuai production yield was increased by 30%, the processing quality was effectively improved, and the effectiveness of the rhuai processing process control was guaranteed. In addition, the results can provide reference for the research and development of intelligent control systems in traditional Chinese medicine preparation industry, and realize the large-scale and industrialized development of traditional Chinese medicine processing equipment.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
With the continuous development and innovation of computer hardware and software, computer communication networks have become an important part of people's daily life and work. Especially all walks of life are constantly referring to cognitive wireless powered communication network (CWPCN). In the application process of CWPCN, it is also necessary to continuously optimize the performance according to the actual application situation, so as to better play the function of CWPCN, only in this way can we bring users a more extreme CWPCN application experience. This paper firstly conducts specific research and analysis on CWPCN deployment and application status, secondly expounds CWPCN performance optimization technology, and lastly summarizes the level of concern users have about network performance and optimization, as well as the areas they think need to be improved in network equipment and software. As a result, CWPCN is playing an increasingly important role in business and personal life, and as such, it needs to be continuously optimized and enhanced to meet user needs.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.