rss_2.0International Journal of Advanced Network, Monitoring and Controls FeedSciendo RSS Feed for International Journal of Advanced Network, Monitoring and Controlshttps://sciendo.com/journal/IJANMChttps://www.sciendo.comInternational Journal of Advanced Network, Monitoring and Controls Feedhttps://sciendo-parsed.s3.eu-central-1.amazonaws.com/660534f21ae47050093cd87e/cover-image.jpghttps://sciendo.com/journal/IJANMC140216Personalized Recommendation Multi-Objective Optimization Model Based on Deep Learninghttps://sciendo.com/article/10.2478/ijanmc-2024-0005<abstract> <title style='display:none'>Abstract</title> <p>Recommended in this paper, because the existing single objective experience is poor, and the recommended model in a large difference of targets under the complex relationship of joint optimization and the conflict caused by faults, this paper proposes a personalized recommendation based on the deep learning multi-objective optimization algorithm, the estimated probability of users on the individual behavior as a model to study target, Multiple objectives are integrated into a model for learning. Firstly, the embedding layer is used to change the feature vectors, so that the bottom layer of the model shares the same feature embedding. Secondly, the factorization machine and deep learning are used to construct high-low order feature interaction. Then, the gating network and multilevel expert network constructed by a fully connected neural network are used to learn the characteristic relationship of user behavior. Finally, make connections between goals. Through experiments on public and real datasets, The results show that the multi-objective model proposed in this paper has better co-optimization performance and increases the AUC value by 0.1% compared with advanced personalized recommendation models such as MMoE and ESMM, to achieve the ultimate goal of increasing the prediction accuracy and improving user satisfaction.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00052024-03-28T00:00:00.000+00:00Face Recognition System Based on Capsule Networkshttps://sciendo.com/article/10.2478/ijanmc-2024-0003<abstract> <title style='display:none'>Abstract</title> <p>This study introduces a technique for facial recognition according to capsule networks. The system utilizes the advantages of capsule networks to model the face features in the image hierarchically, and realizes the efficient recognition of faces. First of all, we know the difference between the capsule network and the convolutional neural network through the study of the operating principle and the structure of the capsule network. Secondly, the Capsule Network is realized through deep research on the algorithm for dynamic routing and the internal operating principle of the capsule. Finally, by conducting experiments on the face dataset and optimizing it with the Adam optimization algorithm as well as the boundary loss and reconstruction loss, the capsule network is promoted to learn more robust feature representations to obtain better face recognition results. The experiments show that the face recognition system based on capsule network can reach 93.5% correct rate of evaluation on WebFace dataset, which achieves a high recognition accuracy. The final results demonstrate the feasibility and effectiveness of capsule networks for face recognition.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00032024-03-28T00:00:00.000+00:00A Target Recognition Method of Small Sample Based on RCS Datahttps://sciendo.com/article/10.2478/ijanmc-2024-0001<abstract> <title style='display:none'>Abstract</title> <p>During the training of target recognition models based on Radar Cross Section (RCS) data, a persistent challenge arises in sampling due to the inherent difficulty in acquiring a sufficient number of samples. This scarcity of data poses a significant impediment to the effective training of models, resulting in diminished accuracy in target recognition. To address this issue, this article proposes a target classification method based on RCS data under small sample conditions. The approach adopts the fundamental concept of Model-Agnostic Meta-Learning (MAML) to train the target recognition model, enhancing the structure of MAML model. An hourglass-shaped convolution layer is introduced to the input layer, with an additional convolution layer preceding the output layer, and a switch to a central loss function. To substantiate the efficacy of the improved MAML model, comprehensive comparative analyses are conducted with benchmark models, including MAML, ResNet 18-layers, Long Short-Term Memory (LSTM), among others. Experimental results conclusively demonstrate the superior performance of the refined MAML model in target recognition under conditions of limited samples, attaining an average prediction accuracy of 85.62%. This signifies a noteworthy 5-percentage-point improvement compared to the baseline model prior to the introduced enhancements.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00012024-03-28T00:00:00.000+00:00Automatic Landing Control of Aircraft Based on Cognitive Load Theory and DDPGhttps://sciendo.com/article/10.2478/ijanmc-2024-0007<abstract> <title style='display:none'>Abstract</title> <p>The keypoint of autonomous driving technology is the accurate instructions maked by desicision-makers based on the perception information. Human plays an important role in the decision-makers. The cognitive load is usually used to quantify the impact of human-computer interaction during flighting. In this paper, we proposed a innovate automatic landing control method based on the cognitive load theory and Deep Deterministic Policy Gradient. Different to the traditional algorithm which heavily relays on an accurate model, the reinforcement learning algorithm is used to design the control strategy in the proposed method. And an improved DDPG algorithm is proposed based on the impact of cognitive load, to improve the training efficiency of the DDPG algorithm and reduce the correlation between data. And construct a human-machine reinforcement learning model. The final position, mean square error of pitch angle, and standard deviation of the aircraft gradually decrease with the number of iterations and tend to 0, indicating that the aircraft is gradually stabilizing its landing. The experimental results demonstrate that the proposed model can greatly improve the longitudinal stability of the aircraft.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00072024-03-28T00:00:00.000+00:00The Time-Sensitive Networking Scheduling Algorithm Based on Q-learninghttps://sciendo.com/article/10.2478/ijanmc-2024-0008<abstract> <title style='display:none'>Abstract</title> <p>Time-Sensitive Networking (TSN) occupies a vital position in modern communication domains, with the 802.1Qbv standard being an important network technology designed to meet real-time requirements. This standard requires network traffic to be transmitted within strict time windows, presenting challenges in network planning, necessitating efficient resource allocation and scheduling strategies. This paper addresses the 802.1Qbv planning problem through the introduction of reinforcement learning algorithms, offering an automated and intelligent solution. We have designed a reinforcement learning agent capable of perceiving network status, learning optimal resource allocation strategies, and dynamically adjusting in real-time environments. Through simulation and experimentation, we have validated the effectiveness of our proposed method, comparing it with traditional planning approaches. The contribution of this study lies in introducing a novel solution to the 802.1Qbv planning problem for time-sensitive networks, enhancing network resource utilization and performance. This approach offers strong support for the development and enhancement of TSN-like networks, holding significant importance for meeting the growing demands of real-time applications.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00082024-03-28T00:00:00.000+00:00Research on Simulation Approximate Solution Strategy for Complex Kinematic Modelshttps://sciendo.com/article/10.2478/ijanmc-2024-0006<abstract> <title style='display:none'>Abstract</title> <p>In order to meet the needs of military, road construction, multimedia industry and other aspects, UAVs are gradually given more functions. As the basic function of UAV applications, the fixed-point delivery problem model has higher and higher accuracy requirements. However, in the actual scene, the UAV delivery problem is affected by the interaction of various factors such as flight height, air resistance, and dive angle, which makes it difficult to achieve high stability and high hit accuracy. This paper will analyze the complex motion model based on the fixed-point delivery of explosives by UAV, study the relationship between the stability of UAV delivery and the hit accuracy, and analyze the influence of relevant parameters on the problem by using modeling. In this paper, a multivariate nonlinear continuous time change model is proposed, and a continuous time slice discretization idea operation model is introduced to approximate the time slice splitting inside the UAV launch motion. Secondly, the design quantified evaluation index reaction the initial velocity of the explosive, the launch Angle, the height off the ground and other parameters to analyze the model. Finally, the best scheduling strategy is obtained and verified by using the idea of variable traversal and trial- and-error simulation. The experimental results show that the variation of UAV flying height, speed, depression and other interference factors is consistent with the prediction of score and hit accuracy, according to the environment setting of this model, when the UAV is 300 meters above the ground and 290 meters away from the target horizontal position, the delivery speed is 250m/s, and the pitch angle is about 27°, the fixed-point delivery of explosives is the best strategy.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00062024-03-28T00:00:00.000+00:00Securing Operating Systems (OS): A Comprehensive Approach to Security with Best Practices and Techniqueshttps://sciendo.com/article/10.2478/ijanmc-2024-0010<abstract> <title style='display:none'>Abstract</title> <p>Operating system (OS) security is paramount in ensuring the integrity, confidentiality, and availability of computer systems and data. This research manuscript presents a comprehensive investigation into the multifaceted domain of OS security, aiming to enhance understanding, identify challenges, and propose effective solutions. The research methodology integrates diverse approaches, including an extensive exploration for available knowledge process mechanics, empirical data collection, case studies investigations, experimental analysis, comparative studies, qualitative analysis, synthesis, and interpretation. Through various experimental perspectives, theoretical foundations, historical developments, and current trends in OS security are also explored. Empirical data collection involves gathering insights from publicly available reports, security advisories, case studies, and expert interviews to capture real-world perspectives and experiences. Case studies illustrate practical implications of security strategies, while experimental analysis evaluates the efficacy of security measures in controlled environments. Comparative studies and qualitative analysis provide insights into strengths, limitations, and emerging trends in OS security. The synthesis and interpretation of the findings offer actionable insights for improving OS security practices, policy recommendations, and providing towards future research directions. This research contributes to advancing knowledge in OS security and informs the development of effective strategies to safeguard computer systems against evolving threats and vulnerabilities.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00102024-03-28T00:00:00.000+00:00A Modified Energy Enhancement in WSN Using the Shortest Path Transmission Techniquehttps://sciendo.com/article/10.2478/ijanmc-2024-0004<abstract> <title style='display:none'>Abstract</title> <p>This study introduced a novel energy enhancement approach for Wireless Sensor Networks (WSNs) by leveraging the shortest path transmission technique to minimize energy consumption and extend the network’s lifetime. Unlike traditional methods that heavily relied on cluster heads (CHs) for data transmission, our model proposed a non-cluster-based routing algorithm, utilizing Dijkstra’s algorithm to identify the most energy-efficient paths for data transmission. Simulation results, based on varying node densities (100, 200, and 300 nodes) within a 200x200 network area, demonstrated the effectiveness of our approach. Our findings indicated a significant reduction in energy consumption, with the network lifetime extending to approximately 100,000 rounds, surpassing traditional LEACH-based and other related protocols. This enhancement not only promised a sustainable WSN deployment but also offered a scalable solution adaptable to different network sizes and configurations.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00042024-03-28T00:00:00.000+00:00Lightweight Low-Altitude UAV Object Detection Based on Improved YOLOv5shttps://sciendo.com/article/10.2478/ijanmc-2024-0009<abstract> <title style='display:none'>Abstract</title> <p>In the context of rapid developments in drone technology, the significance of recognizing and detecting low-altitude unmanned aerial vehicles (UAVs) has grown. Although conventional algorithmic enhancements have increased the detection rate of low-altitude UAV targets, they tend to neglect the intricate nature and computational demands of the algorithms. This paper introduces ATD-YOLO, an enhanced target detection model based on the YOLOv5s architecture, aimed at tackling this issue. Firstly, a realistic low-altitude UAV dataset is fashioned by amalgamating various publicly available datasets. Secondly, a C3F module grounded in FasterNet, incorporating Partial Convolution (PConv), is introduced to decrease model parameters while upholding detection accuracy. Furthermore, the backbone network incorporates an Efficient Multi-Scale Attention (EMA) module to extract essential image information while filtering out irrelevant details, facilitating adaptive feature fusion. Additionally, the universal upsampling operator CARAFE (Content-aware reassembly of features) is utilized instead of nearest-neighbor upsampling. This enhancement boosts the performance of the feature pyramid network by expanding the receptive field for data feature fusion. Lastly, the Slim-Neck network is introduced to fine-tune the feature fusion network, thereby reducing the model’s floating-point calculations and parameters. Experimental findings demonstrate that the improved ATD-YOLO model achieves an accuracy of 92.8%, with a 31.4% decrease in parameters and a 28.7% decrease in floating-point calculations compared to the original model. The detection speed reaches 75.37 frames per second (FPS). These experiments affirm that the proposed enhancement method meets the deployment requirements for low computational power while maintaining high precision.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00092024-03-28T00:00:00.000+00:00Indoor Robot SLAM with Multi-Sensor Fusionhttps://sciendo.com/article/10.2478/ijanmc-2024-0002<abstract> <title style='display:none'>Abstract</title> <p>In order to solve the problem of large positioning error and incomplete mapping of SLAM based on two-dimensional lidar in indoor environment, a multi-sensor fusion SLAM algorithm for indoor robots was proposed. Aiming at the mismatch problem of the traditional ICP algorithm in the front end of the lidar SLAM, the algorithm adopts the PL-ICP algorithm that is more suitable for the indoor environment, and uses the extended Kalman filter to fuse the wheel odometer and IMU to provide the initial motion estimation value. Then, during the mapping phase, the pseudo 2D laser data converted from the 3D point cloud data obtained by the depth camera is fused with the data obtained from the 2D lidar to compensate for the lack of vertical field of view in the 2D lidar mapping. The final experimental results show that the fusion odometer data has improved the positioning accuracy by at least 33% compared to a single wheeled odometer, providing a higher initial iteration value for the PL-ICP algorithm. At the same time, fusion mapping compensates for the shortcomings of a single two-dimensional lidar mapping, and constructs an environmental map with more complete environmental information.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2024-00022024-03-28T00:00:00.000+00:00GreatFree as a Generic Distributed Programming Language and the Foundation of the Cloud-Side Operating Systemhttps://sciendo.com/article/10.2478/ijanmc-2023-0078<abstract> <title style='display:none'>Abstract</title> <p>GreatFree is a generic distributed programming language to develop various distributed systems over the Internet-oriented computing environment. The fundamental characters of GreatFree are shaped by three essential techniques, including the message-passing, the physical-machine-visible, and the thread-visible. More important, GreatFree is equipped with three additional distinguished mechanisms, i.e., the distributed primitives, the distributed common patterns, and the distributed threads on the application level, which are sufficient to turn GreatFree into a generic distributed programming technology. To the best of our knowledge, compared with any others, GreatFree is the first one to achieve the goal. Thereafter, GreatFree is capable of exploiting distributed computing resources flexibly to adapt to any heterogeneous environments with a uniform solution. It indicates that GreatFree represents the common principles existed in various complicated computing circumstances over the Internet. That inspires that GreatFree is a proper technology to build a new concept of cloud computing environment, i.e., the cloud-side operating system, which dominates diverse distributed computing resources upon the common principles of GreatFree. Such a system is a generic development and running environment for any distributed systems. Without doubt, within the environment, GreatFree is the unique choice to program any distributed systems in a scalable manner.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00782024-03-16T00:00:00.000+00:00Design and Implementation of Smart Home System Based on STM32 Microcomputerhttps://sciendo.com/article/10.2478/ijanmc-2023-0072<abstract> <title style='display:none'>Abstract</title> <p>With the rapid development of the Internet of Things science and technology, people′s living standards are gradually improving, and the requirements for the living environment are also getting higher and higher, which makes smart homes gradually enter thousands of households. The purpose of this project is to design a system that integrates hardware and software and can measure and transmit various data. Among them, the hardware part includes data measurement and data display. The data measurement module consists of DHT11 temperature and humidity sensor, DSM501 particle number sensor and MQ3 alcohol concentration sensor. The experimental data will be displayed on the TFTLCD screen. The system software is partly run on the Windows operating system, using the Python language development. This system takes ESP8266 module as the transfer station, realizes the communication between STM32 development board and computer. The experiment shows that the system has the advantages of high measurement data accuracy, fast data refresh speed, complete data transmission, simple design, high reliability, easy installation, economical and practical, and has certain practical value in life, production, industry and other fields.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00722024-03-16T00:00:00.000+00:00The Application of Whale Optimization Algorithm in Array Antennashttps://sciendo.com/article/10.2478/ijanmc-2023-0074<abstract> <title style='display:none'>Abstract</title> <p>With the continuous improvement of various radio system performance indicators, the research work on antenna has become particularly important. According to different scenarios and requirements, practical projects also need the corresponding antennas to produce different radiation patterns. By reasonably setting the parameters of the array antenna, the target radiation pattern can be obtained to meet real life applications. When the array antenna has a large number of basic units and the expected far-field pattern is complicated, the design of the array antenna becomes a complicated optimization problem. To solve this problem, Whale Optimization Algorithm (WOA) is proposed. WOA is not only simple and fast, but can also get the global optimal solution. Therefore, WOA has developed rapidly in recent years. However, the application of this algorithm in the field of antenna design is still relatively rare, thus using WOA to solve the optimization problem of array antenna design is very valuable.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00742024-03-16T00:00:00.000+00:00Research on Multi-Person Pose Estimation Technologyhttps://sciendo.com/article/10.2478/ijanmc-2023-0075<abstract> <title style='display:none'>Abstract</title> <p>Human pose estimation is a hot topic of computer research in recent years, which promotes the progress of society and brings many conveniences to people’s lives. Fom traditional methods to the mainstream deep learning-based methods, the primary approach in deep learning involves the use of convolutional neural networks to reduce computational complexity and improve network accuracy, but because the network structure is too deep to improve the accuracy, the trained model parameters are also very large, and it is very dependent on the input of hardware equipment. At this time, the lightweight human pose estimation can solve this problem very well. This paper mainly describes the knowledge of convolutional neural network in detail and compares it with traditional image algorithms. The OpenPose model is a classic model based on convolutional neural network that can well achieve single-person and multi-person human pose estimation model, but because the convolution kernel in its network structure is too large to increase the amount of calculation, this paper proposes three improvements to the network structure of the conventional OpenPose model. Finally, the precision of the model is improved by about 40%, which verifies the feasibility of lightweight human body posture estimation research.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00752024-03-16T00:00:00.000+00:00Research on Object Detection in Animal Images Based on Convolutional Neural Networkshttps://sciendo.com/article/10.2478/ijanmc-2023-0076<abstract> <title style='display:none'>Abstract</title> <p>Object detection is the use of computer to find out all the objects of interest in the image, determine their categories and locations, is one of the core problems in the field of computer vision. Traditional animal image target detection usually adopts the sliding window method, but due to the different sizes of the input images, this method has some problems such as insufficient training samples, low detection accuracy and slow speed. In order to solve such problems, based on the development of deep learning in recent years, this paper proposes an object detection algorithm based on convolutional neural network. YOLOv5 was used to effectively distinguish, identify and mark animal categories, which accelerated the training of the model and greatly improved the accuracy of target detection. Through the analysis of experimental data, it was concluded that the algorithm studied in this paper had good performance and good target detection results. Finally, the key problems of object detection research are summarized, and the future development trend of this field is prospected. When the number of training rounds is 30, the accuracy rate has reached about 70%, and after 50 rounds of training, some accuracy can reach 90%, which is excellent and better than other traditional algorithms.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00762024-03-16T00:00:00.000+00:00Deep Learning Based Recognition of Lepidoptera Insectshttps://sciendo.com/article/10.2478/ijanmc-2023-0073<abstract> <title style='display:none'>Abstract</title> <p>The successful application of cutting-edge computer vision technology to automatic insect classification has long been a focus of research in insect taxonomy. The results of this research have a wide range of applications in areas such as environmental monitoring, pest diagnosis and epidemiology. However, there is still a gap between the current techniques used in automatic insect classification and the latest computer vision techniques. The research in this paper is conducted on Lepidoptera, a class of insects that are widely infested, including butterflies and moths. The study focuses on the application of deep learning algorithms in image processing of Lepidoptera insects. In order to improve the recognition rate for Lepidoptera insect recognition, this paper uses a detection model based on deep neural networks to realize the recognition of Lepidoptera insects in complex environments. Specifically, the yolov7 algorithm is adopted as the basic model for this experiment, and the reasons for using this model are explained in terms of the splicing of network modules, loss function, positive sample allocation strategy, and the merging of convolution and normalization, respectively. Through experiments, it is proved that the algorithm can effectively improve the gesture recognition rate, the recognition accuracy reaches 79.5%, and the recognition speed is as high as 33.08it/s.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00732024-03-16T00:00:00.000+00:00A Model-Based Approach to Mobile Application Testinghttps://sciendo.com/article/10.2478/ijanmc-2023-0071<abstract> <title style='display:none'>Abstract</title> <p>Modeling the automated testing of mobile applications is a crucial aspect of mobile application automation testing. Due to the varied styles and complex interactions of mobile applications, automated modeling methods are urgently required, particularly in the context of their short development cycles, large numbers, and fast version iterations and updates. In this paper, we address the challenge of exploring mobile application behavior and state based on robotic testing environment without invading the application interior, and propose a method for automated exploration of GUI components and GUI events of applications combined with application domain knowledge to generate mobile application GUI semantic test models. Our results show that the proposed semantic model achieves 70.6% and 82.4% defect detection rate in the robot vision environment and simulation environment, respectively. Compared with the comparative testing method that can only find application crash defects, our method can explore both crash defects and functional anomalies with the application semantic understanding and domain knowledge, thereby extending the automated mobile application functional testing capability of mobile applications. In response to the limitations of mobile application automated testing modeling mentioned above, this paper introduces an automated testing method based on semantic models. It uses the proposed semantic testing model to guide the purposeful exploration of the tested application’s states. Subsequently, it generates positive and negative test cases based on the domain knowledge associated with the semantic model. This modeling approach leverages domain models in the mobile application field to conduct automated modeling tests imbued with functional significance, guided by domain knowledge. This optimization aims to address the shortcomings of current automated testing, particularly in terms of model reuse and test expansion.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00712024-03-16T00:00:00.000+00:00Research and Implementation of Forest Fire Detection Algorithm Improvementhttps://sciendo.com/article/10.2478/ijanmc-2023-0080<abstract> <title style='display:none'>Abstract</title> <p>To overcome low efficiency and accuracy of existing forest fire detection algorithms, this paper proposes a network model to enhance the real-time and robustness of detection. This structure is based on the YOLOv5 target detection algorithm and combines the backbone network with The feature extraction module combines the attention module dsCBAM improved by depth-separable convolution, and replaces the loss function CIoU of the original model with a VariFocal loss function that is more suitable for the imbalanced characteristics of positive and negative samples in the forest fire data set. Experiments were conducted on a self-made and public forest fire data set. The accuracy and recall rate of the model can reach 87.1% and 81.6%, which are 7.40% and 3.20% higher than the original model, and the number of images processed per second reaches 64 frames, a growth rate of 8.47%. At the same time, this model was compared horizontally with other improved methods. The accuracy, recall rate and processing speed were all improved in the range of 3% to 10%. The effectiveness of the improved method in this article was verified, and the external perception level of the forest fire scene was deeper.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00802024-03-16T00:00:00.000+00:00Object Localization Algorithm Based on Meta-Reinforcement Learninghttps://sciendo.com/article/10.2478/ijanmc-2023-0077<abstract> <title style='display:none'>Abstract</title> <p>When the target localization algorithm based on reinforcement learning is trained on few-sample data sets, the accuracy of target localization is low due to the low degree of fitting. Therefore, on the basis of deep reinforcement learning target localization algorithm, this paper proposes a target localization algorithm based on meta-reinforcement learning. Firstly, during the initial training of the model, the meta-parameters were classified and stored according to the similarity of the training tasks. Then, for the new target location task, the task feature extraction was carried out and the meta parameters with the highest similarity were matched as the initial parameters of the model training. The model dynamically updated the meta parameter pool to ensure that the optimal meta parameters of multiple different types of features were saved in the meta parameter pool, so as to improve the generalization ability and recognition accuracy of multiple types of target location tasks. Experimental results show that in a variety of single target localization tasks, compared with the original reinforcement learning target localization algorithm, under the same data set size, the model converges under a small number of training steps with the meta-parameters in the matching meta-parameter pool as the initial training parameters. Moreover, the training speed of the meta-reinforcement learning method based on MAML-RL is increased by 28.2% for random initial parameters, and that of the meta-reinforcement learning method based on this paper is increased by 34.9%, indicating that the proposed algorithm effectively improves the training speed, generalization performance and localization accuracy of object detection.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00772024-03-16T00:00:00.000+00:00Research on Improved Dual Channel Medical Short Text Intention Recognition Algorithmhttps://sciendo.com/article/10.2478/ijanmc-2023-0079<abstract> <title style='display:none'>Abstract</title> <p>The increasing application of medical robots in the healthcare sector underscores the critical importance of intent recognition in enhancing the interaction and assistance capabilities of these robots. Traditional intent recognition methods utilize convolutional neural networks (CNNs) for text analysis but often fall short in capturing global features, resulting in incomplete information. To address this challenge, this paper introduces an innovative approach by combining an enhanced CNN with bidirectional gated recurrent units (BiGRU) to construct a dual-channel short-text intent recognition model. This model effectively leverages both local and global features to more accurately comprehend user needs and intentions. Experimental results demonstrate that this model excels, achieving an accuracy rate of 96.68% and an F1 score of 96.67% on the THUCNews_Title dataset. In comparison to conventional intent recognition models, it exhibits significantly improved performance, thereby providing substantial support for medical robots in patient care and assisting healthcare professionals.</p> </abstract>ARTICLEtruehttps://sciendo.com/article/10.2478/ijanmc-2023-00792024-03-16T00:00:00.000+00:00en-us-1