Salah-Eddine Mansour, Abdelhak Sakhi, Larbi Kzaz, A. Sekkaki
This study contributes to improving Morocco's fish canning industry by integrating artificial intelligence (AI). The primary objective involves developing an AI and image processing-based system to monitor and guarantee canning process quality in the facility. It commenced with an IoT-enabled device capable of capturing and processing images, leading to the creation of an AI-driven system adept at accurately categorizing improperly crimped cans. Further advancements focused on reinforcing communication between IoT devices and servers housing individual client's neural network weights. These weights are vital, ensuring the functionality of our IoT device. The efficiency of the IoT device in categorizing cans relies on updated neural network weights from the Fog server, crucial for continual refinement and adaptation to diverse can shapes. Securing communication integrity between devices and the server is imperative to avoid disruptions in can classification, emphasizing the need for secure channels. In this paper, our key scientific contribution revolves around devising a security protocol founded on HMAC. This protocol guarantees authentication and preserves the integrity of neural network weights exchanged between Fog computing nodes and IoT devices. The innovative addition of a comprehensive dictionary within the Fog server significantly bolsters security measures, enhancing the overall safety between these interconnected entities.
{"title":"Enhancing Security Mechanisms for IoT-Fog Networks","authors":"Salah-Eddine Mansour, Abdelhak Sakhi, Larbi Kzaz, A. Sekkaki","doi":"10.18196/jrc.v5i1.20745","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20745","url":null,"abstract":"This study contributes to improving Morocco's fish canning industry by integrating artificial intelligence (AI). The primary objective involves developing an AI and image processing-based system to monitor and guarantee canning process quality in the facility. It commenced with an IoT-enabled device capable of capturing and processing images, leading to the creation of an AI-driven system adept at accurately categorizing improperly crimped cans. Further advancements focused on reinforcing communication between IoT devices and servers housing individual client's neural network weights. These weights are vital, ensuring the functionality of our IoT device. The efficiency of the IoT device in categorizing cans relies on updated neural network weights from the Fog server, crucial for continual refinement and adaptation to diverse can shapes. Securing communication integrity between devices and the server is imperative to avoid disruptions in can classification, emphasizing the need for secure channels. In this paper, our key scientific contribution revolves around devising a security protocol founded on HMAC. This protocol guarantees authentication and preserves the integrity of neural network weights exchanged between Fog computing nodes and IoT devices. The innovative addition of a comprehensive dictionary within the Fog server significantly bolsters security measures, enhancing the overall safety between these interconnected entities.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"51 5","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140497368","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This research investigates the impact of model simplification on the dynamic performance of an ABB IRB-140 six-jointed industrial robotic arm, concentrating on torque prediction and energy consumption. The entire mathematical model of forward, reverse, differential kinematics, and dynamic model proposed based on the technical specifications of the arm, and to obtain the center of the mass and inertia matrices, which are essential components of the dynamic model, Utilizing Solidworks, we developed three CAD/CAM models representing the manipulator with varying detail levels, such as simplified, semi-detailed, and detailed. Our findings indicate minor differences in the model's torque and energy consumption graphs. The semi-detailed model consumed the most energy, except for joint 1, with the detailed model showing a 0.53% reduction and the simplified model a 6.8% reduction in energy consumption. Despite these variations, all models proved effective in predicting the robot's performance during a standard 30-second task, demonstrating their adequacy for various industrial applications. This research highlights the balance between computational efficiency and accuracy in model selection. While the detailed model offers the highest precision, it demands more computational resources, which is suitable for high-precision tasks. In discrepancy, simplified, less precise models offer computational efficiency, making them adequate for specific scenarios. Our study provides critical insights into selecting dynamic models in industrial robotics. It guides the optimization of performance and energy efficiency based on the required task precision and available computational resources. This comprehensive comparison of dynamic models underscores their applicability and effectiveness in diverse industrial settings.
{"title":"The Impact of Simplifications of the Dynamic Model on the Motion of a Six-Jointed Industrial Articulated Robotic Arm Movement","authors":"Mehdi Fazilat, N. Zioui","doi":"10.18196/jrc.v5i1.20263","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20263","url":null,"abstract":"This research investigates the impact of model simplification on the dynamic performance of an ABB IRB-140 six-jointed industrial robotic arm, concentrating on torque prediction and energy consumption. The entire mathematical model of forward, reverse, differential kinematics, and dynamic model proposed based on the technical specifications of the arm, and to obtain the center of the mass and inertia matrices, which are essential components of the dynamic model, Utilizing Solidworks, we developed three CAD/CAM models representing the manipulator with varying detail levels, such as simplified, semi-detailed, and detailed. Our findings indicate minor differences in the model's torque and energy consumption graphs. The semi-detailed model consumed the most energy, except for joint 1, with the detailed model showing a 0.53% reduction and the simplified model a 6.8% reduction in energy consumption. Despite these variations, all models proved effective in predicting the robot's performance during a standard 30-second task, demonstrating their adequacy for various industrial applications. This research highlights the balance between computational efficiency and accuracy in model selection. While the detailed model offers the highest precision, it demands more computational resources, which is suitable for high-precision tasks. In discrepancy, simplified, less precise models offer computational efficiency, making them adequate for specific scenarios. Our study provides critical insights into selecting dynamic models in industrial robotics. It guides the optimization of performance and energy efficiency based on the required task precision and available computational resources. This comprehensive comparison of dynamic models underscores their applicability and effectiveness in diverse industrial settings.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"55 3","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140497555","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Mohammad Afkar, R. Gavagsaz-Ghoachani, M. Phattanasak, Serge Pierfederici, Wiset Saksiri
Switching power converters are integral in various applications like transportation and renewable energy. After their design, ensuring stable closed-loop poles is critical to maintain safe operating conditions. This study focuses on a switching DC-DC boost converter with a cascade control approach using an energy controller for the outer loop and indirect-sliding mode control for the inner loop. The research objective involves investigating stability through eigenvalue evaluation at different operating points. A large-signal average model is applied to make controlled performance independent of the operating point by fixing system poles. Nonlinear controllers, specifically indirect-sliding mode control, are chosen for their robustness, constant switching frequency, and implementation ease. Results indicate that insufficient decoupling leads to eigenvalue displacement, impacting control parameter choices. The research contribution is investigating the local stability of cascaded control, considering its advantageous implications for both performance and design. This study contributes to the understanding of switching power converters' stability, emphasizing the proposed methodology's broader applicability to diverse converter structures. The proposed approach, applicable to various switching power converters, sheds light on the importance of proper decoupling between outer and inner loop dynamics.
{"title":"Local-Stability Analysis of Cascaded Control for a Switching Power Converter","authors":"Mohammad Afkar, R. Gavagsaz-Ghoachani, M. Phattanasak, Serge Pierfederici, Wiset Saksiri","doi":"10.18196/jrc.v5i1.20302","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20302","url":null,"abstract":"Switching power converters are integral in various applications like transportation and renewable energy. After their design, ensuring stable closed-loop poles is critical to maintain safe operating conditions. This study focuses on a switching DC-DC boost converter with a cascade control approach using an energy controller for the outer loop and indirect-sliding mode control for the inner loop. The research objective involves investigating stability through eigenvalue evaluation at different operating points. A large-signal average model is applied to make controlled performance independent of the operating point by fixing system poles. Nonlinear controllers, specifically indirect-sliding mode control, are chosen for their robustness, constant switching frequency, and implementation ease. Results indicate that insufficient decoupling leads to eigenvalue displacement, impacting control parameter choices. The research contribution is investigating the local stability of cascaded control, considering its advantageous implications for both performance and design. This study contributes to the understanding of switching power converters' stability, emphasizing the proposed methodology's broader applicability to diverse converter structures. The proposed approach, applicable to various switching power converters, sheds light on the importance of proper decoupling between outer and inner loop dynamics.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"7 2","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140497755","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Autonomous mobile robots based on wheel drive are widely used in various applications. The differential drive mobile robot (DDMR) is one type with wheel drive. DDMR uses one actuator to move each wheel on the mobile robot. Autonomous capabilities are needed to avoid obstacles around the DDMR. This paper presents implementing a fuzzy logic algorithm for obstacle avoidance at a low cost (DDMR). The fuzzy logic algorithm input is obtained from three ultrasonic sensors installed in front of the DDMR with an angle difference between the sensors of 45$^0$. Distance information from the ultrasonic sensors is used to regulate the speed of the right and left motors of the DDMR. Based on the test results, the Mamdani inference system using the fuzzy logic algorithm was successfully implemented as an obstacle avoidance algorithm. The speed values of the right and left DDMR wheels produce values according to the rules created in the Mamdani inference system. DDMR managed to pass through a tunnel-shaped environment and reach its goal without hitting any obstacles around it. The average speed produced by DDMR in reaching the goal is 4.91 cm/s.
{"title":"Design and Implementation of Fuzzy Logic for Obstacle Avoidance in Differential Drive Mobile Robot","authors":"R. Puriyanto, Ahmad Kamal Mustofa","doi":"10.18196/jrc.v5i1.20524","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20524","url":null,"abstract":"Autonomous mobile robots based on wheel drive are widely used in various applications. The differential drive mobile robot (DDMR) is one type with wheel drive. DDMR uses one actuator to move each wheel on the mobile robot. Autonomous capabilities are needed to avoid obstacles around the DDMR. This paper presents implementing a fuzzy logic algorithm for obstacle avoidance at a low cost (DDMR). The fuzzy logic algorithm input is obtained from three ultrasonic sensors installed in front of the DDMR with an angle difference between the sensors of 45$^0$. Distance information from the ultrasonic sensors is used to regulate the speed of the right and left motors of the DDMR. Based on the test results, the Mamdani inference system using the fuzzy logic algorithm was successfully implemented as an obstacle avoidance algorithm. The speed values of the right and left DDMR wheels produce values according to the rules created in the Mamdani inference system. DDMR managed to pass through a tunnel-shaped environment and reach its goal without hitting any obstacles around it. The average speed produced by DDMR in reaching the goal is 4.91 cm/s.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"47 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140507890","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Omar Y. Ismael, Mohammed Almaged, Abdulla Ibrahim Abdulla
This work proposes an efficient and safe single-layer Nonlinear Model Predictive Control (NMPC) system based on LiDAR to solve the problem of autonomous navigation in cluttered environments with previously unidentified static and dynamic obstacles of any shape. Initially, LiDAR sensor data is collected. Then, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, is used to cluster the (Lidar) points that belong to each obstacle together. Moreover, a Minimum Euclidean Distance (MED) between the robot and each obstacle with the aid of a safety margin is utilized to implement safety-critical obstacle avoidance rather than existing methods in the literature that depend on enclosing the obstacles with a circle or minimum bounding ellipse. After that, to impose avoidance constraints with feasibility guarantees and without compromising stability, an NMPC for set-point stabilization is taken into consideration with a design strategy based on terminal inequality and equality constraints. Consequently, numerous obstacles can be avoided at the same time efficiently and rapidly through unstructured environments with narrow corridors. Finally, a case study with an omnidirectional wheeled mobile robot (OWMR) is presented to assess the proposed NMPC formulation for set-point stabilization. Furthermore, the efficacy of the proposed system is tested by experiments in simulated scenarios using a robot simulator named CoppeliaSim in combination with MATLAB which utilizes the CasADi Toolbox, and Statistics and Machine Learning Toolbox. Two simulation scenarios are considered to show the performance of the proposed framework. The first scenario considers only static obstacles while the second scenario is more challenging and contains static and dynamic obstacles. In both scenarios, the OWMR successfully reached the target pose (1.5m, 1.5m, 0°) with a small deviation. Four performance indices are utilized to evaluate the set-point stabilization performance of the proposed control framework including the steady-state error in the posture vector which is less than 0.02 meters for position and 0.012 for orientation, and the integral of norm squared actual control inputs which is 19.96 and 21.74 for the first and second scenarios respectively. The proposed control framework shows a positive performance in a narrow-cluttered environment with unknown obstacles.
{"title":"Nonlinear Model Predictive Control-based Collision Avoidance for Mobile Robot","authors":"Omar Y. Ismael, Mohammed Almaged, Abdulla Ibrahim Abdulla","doi":"10.18196/jrc.v5i1.20615","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20615","url":null,"abstract":"This work proposes an efficient and safe single-layer Nonlinear Model Predictive Control (NMPC) system based on LiDAR to solve the problem of autonomous navigation in cluttered environments with previously unidentified static and dynamic obstacles of any shape. Initially, LiDAR sensor data is collected. Then, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, is used to cluster the (Lidar) points that belong to each obstacle together. Moreover, a Minimum Euclidean Distance (MED) between the robot and each obstacle with the aid of a safety margin is utilized to implement safety-critical obstacle avoidance rather than existing methods in the literature that depend on enclosing the obstacles with a circle or minimum bounding ellipse. After that, to impose avoidance constraints with feasibility guarantees and without compromising stability, an NMPC for set-point stabilization is taken into consideration with a design strategy based on terminal inequality and equality constraints. Consequently, numerous obstacles can be avoided at the same time efficiently and rapidly through unstructured environments with narrow corridors. Finally, a case study with an omnidirectional wheeled mobile robot (OWMR) is presented to assess the proposed NMPC formulation for set-point stabilization. Furthermore, the efficacy of the proposed system is tested by experiments in simulated scenarios using a robot simulator named CoppeliaSim in combination with MATLAB which utilizes the CasADi Toolbox, and Statistics and Machine Learning Toolbox. Two simulation scenarios are considered to show the performance of the proposed framework. The first scenario considers only static obstacles while the second scenario is more challenging and contains static and dynamic obstacles. In both scenarios, the OWMR successfully reached the target pose (1.5m, 1.5m, 0°) with a small deviation. Four performance indices are utilized to evaluate the set-point stabilization performance of the proposed control framework including the steady-state error in the posture vector which is less than 0.02 meters for position and 0.012 for orientation, and the integral of norm squared actual control inputs which is 19.96 and 21.74 for the first and second scenarios respectively. The proposed control framework shows a positive performance in a narrow-cluttered environment with unknown obstacles.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"10 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140507128","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper, we focus on enhancing the performance of the controller utilized in the Stewart platform by investigating the dynamics of the platform. Dynamic modeling is crucial for control and simulation, yet challenging for parallel robots like the Stewart platform due to closed-loop kinematics. We explore classical methods to solve its inverse dynamical model, but conventional approaches face difficulties, often resulting in simplified and inaccurate models. To overcome this limitation, we propose a novel approach by replacing the classical feedforward inverse dynamic block with a reinforcement learning (RL) agent, which, to our knowledge, has not been tried yet in the context of the Stewart platform control. Our proposed methodology utilizes a hybrid control topology that combines RL with existing classical control topologies and inverse kinematic modeling. We leverage three deep reinforcement learning (DRL) algorithms and two model-based RL algorithms to achieve improved control performance, highlighting the versatility of the proposed approach. By incorporating the learned feedforward control topology into the existing PID controller, we demonstrate enhancements in the overall control performance of the Stewart platform. Notably, our approach eliminates the need for explicit derivation and solving of the inverse dynamic model, overcoming the drawbacks associated with inaccurate and simplified models. Through several simulations and experiments, we validate the effectiveness of our reinforcement learning-based control approach for the dynamic modeling of the Stewart platform. The results highlight the potential of RL techniques in overcoming the challenges associated with dynamic modeling in parallel robot systems, promising improved control performance. This enhances accuracy and reduces the development time of control algorithms in real-world applications. Nonetheless, it requires a simulation step before practical implementations.
在本文中，我们将通过研究 Stewart 平台的动态特性，重点提高该平台所使用控制器的性能。动态建模对于控制和仿真至关重要，但对于像 Stewart 平台这样的并联机器人来说，由于其闭环运动学特性，动态建模具有挑战性。我们探索了经典方法来求解其逆动力学模型，但传统方法面临着困难，往往会导致模型的简化和不准确。为了克服这一局限，我们提出了一种新方法，即用强化学习（RL）代理取代经典的前馈逆动态模块。我们提出的方法利用混合控制拓扑，将 RL 与现有的经典控制拓扑和逆运动学建模相结合。我们利用三种深度强化学习（DRL）算法和两种基于模型的 RL 算法来提高控制性能，从而凸显了所提方法的多功能性。通过将学习到的前馈控制拓扑纳入现有的 PID 控制器，我们展示了 Stewart 平台整体控制性能的提升。值得注意的是，我们的方法无需明确推导和求解逆动态模型，克服了与不准确和简化模型相关的缺点。通过多次模拟和实验，我们验证了基于强化学习的控制方法对 Stewart 平台动态建模的有效性。结果凸显了强化学习技术在克服并行机器人系统动态建模相关挑战方面的潜力，有望改善控制性能。这提高了控制算法在实际应用中的准确性并缩短了开发时间。不过，在实际应用之前，还需要进行仿真。
{"title":"Addressing Challenges in Dynamic Modeling of Stewart Platform using Reinforcement Learning-Based Control Approach","authors":"H. Yadavari, Vahid Tavakol Aghaei, S. Ikizoglu","doi":"10.18196/jrc.v5i1.20582","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20582","url":null,"abstract":"In this paper, we focus on enhancing the performance of the controller utilized in the Stewart platform by investigating the dynamics of the platform. Dynamic modeling is crucial for control and simulation, yet challenging for parallel robots like the Stewart platform due to closed-loop kinematics. We explore classical methods to solve its inverse dynamical model, but conventional approaches face difficulties, often resulting in simplified and inaccurate models. To overcome this limitation, we propose a novel approach by replacing the classical feedforward inverse dynamic block with a reinforcement learning (RL) agent, which, to our knowledge, has not been tried yet in the context of the Stewart platform control. Our proposed methodology utilizes a hybrid control topology that combines RL with existing classical control topologies and inverse kinematic modeling. We leverage three deep reinforcement learning (DRL) algorithms and two model-based RL algorithms to achieve improved control performance, highlighting the versatility of the proposed approach. By incorporating the learned feedforward control topology into the existing PID controller, we demonstrate enhancements in the overall control performance of the Stewart platform. Notably, our approach eliminates the need for explicit derivation and solving of the inverse dynamic model, overcoming the drawbacks associated with inaccurate and simplified models. Through several simulations and experiments, we validate the effectiveness of our reinforcement learning-based control approach for the dynamic modeling of the Stewart platform. The results highlight the potential of RL techniques in overcoming the challenges associated with dynamic modeling in parallel robot systems, promising improved control performance. This enhances accuracy and reduces the development time of control algorithms in real-world applications. Nonetheless, it requires a simulation step before practical implementations.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"8 4","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140509340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Discriminative feature embedding is used for largescale facial recognition. Many image-based facial recognition networks use CNNs like ResNets and VGG-nets. Humans prioritise different elements, but CNNs treat all facial pictures equally. NLP and computer vision use attention to learn the most important part of an input signal. The inter-channel and inter-spatial attention mechanism is used to assess face image component significance in this study. Channel scalars are calculated using Global Average Pooling in face recognition channel attention. A recent study found that GAP encodes low-frequency channel information first. We compressed channels using discrete cosine transform (DCT) instead of scalar representation to evaluate information at frequencies other than the lowest frequency for the channel attention mechanism. Later layers can acquire the feature map after spatial attention. Channel and spatial attention increase CNN facial recognition feature extraction. Channel-only, spatial-only, parallel, sequential, or channel-after-spatial attention blocks exist. Current face recognition attention approaches may be outperformed on public datasets (Labelled Faces in the Wild).
{"title":"Leveraging a Two-Level Attention Mechanism for Deep Face Recognition with Siamese One-Shot Learning","authors":"Arkan Mahmood Albayati, Wael Chtourou, F. Zarai","doi":"10.18196/jrc.v5i1.20135","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20135","url":null,"abstract":"Discriminative feature embedding is used for largescale facial recognition. Many image-based facial recognition networks use CNNs like ResNets and VGG-nets. Humans prioritise different elements, but CNNs treat all facial pictures equally. NLP and computer vision use attention to learn the most important part of an input signal. The inter-channel and inter-spatial attention mechanism is used to assess face image component significance in this study. Channel scalars are calculated using Global Average Pooling in face recognition channel attention. A recent study found that GAP encodes low-frequency channel information first. We compressed channels using discrete cosine transform (DCT) instead of scalar representation to evaluate information at frequencies other than the lowest frequency for the channel attention mechanism. Later layers can acquire the feature map after spatial attention. Channel and spatial attention increase CNN facial recognition feature extraction. Channel-only, spatial-only, parallel, sequential, or channel-after-spatial attention blocks exist. Current face recognition attention approaches may be outperformed on public datasets (Labelled Faces in the Wild).","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"15 7","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140510186","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Muhammad Syukri Sadimantara, B. D. Argo, Sucipto Sucipto, D. F. Al Riza, Yusuf Hendrawan
Aflatoxin contamination in cacao is a significant problem in terms of trade losses and health effects. This calls for the need for a non-invasive, precise, and effective detection strategy. This research contribution is to determine the best deep-learning model to classify the aflatoxin contamination level in cocoa beans based on fluorescence images and deep learning to improve performance in the classification. The process involved inoculating and incubating Aspergillus flavus (6mL/100g) to obtain aflatoxin-contaminated cocoa beans for 7 days during the incubation period. Liquid Mass Chromatography (LCMS) was used to quantify the aflatoxin in order to categorize the images into different levels including “free of aflatoxin”, “contaminated below the limit”, and “contaminated above the limit”. 300 images were acquired through a mini studio equipped with UV lamps. The aflatoxin level was classified using several pre-trained CNN approaches which has high accuracy such as GoogLeNet, SqueezeNet, AlexNet, and ResNet50. The sensitivity analysis showed that the highest classification accuracy was found in the GoogLeNet model with optimizer: Adam and learning rate: 0.0001 by 96.42%. The model was tested using a testing dataset and obtain accuracy of 96% based on the confusion matrix. The findings indicate that combining CNN with fluorescence images improved the ability to classify the amount of aflatoxin contamination in cacao beans. This method has the potential to be more accurate and economical than the current approach, which could be adapted to reduce aflatoxin's negative effects on food safety and cacao trade losses.
{"title":"The Classification of Aflatoxin Contamination Level in Cocoa Beans using Fluorescence Imaging and Deep learning","authors":"Muhammad Syukri Sadimantara, B. D. Argo, Sucipto Sucipto, D. F. Al Riza, Yusuf Hendrawan","doi":"10.18196/jrc.v5i1.19081","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.19081","url":null,"abstract":"Aflatoxin contamination in cacao is a significant problem in terms of trade losses and health effects. This calls for the need for a non-invasive, precise, and effective detection strategy. This research contribution is to determine the best deep-learning model to classify the aflatoxin contamination level in cocoa beans based on fluorescence images and deep learning to improve performance in the classification. The process involved inoculating and incubating Aspergillus flavus (6mL/100g) to obtain aflatoxin-contaminated cocoa beans for 7 days during the incubation period. Liquid Mass Chromatography (LCMS) was used to quantify the aflatoxin in order to categorize the images into different levels including “free of aflatoxin”, “contaminated below the limit”, and “contaminated above the limit”. 300 images were acquired through a mini studio equipped with UV lamps. The aflatoxin level was classified using several pre-trained CNN approaches which has high accuracy such as GoogLeNet, SqueezeNet, AlexNet, and ResNet50. The sensitivity analysis showed that the highest classification accuracy was found in the GoogLeNet model with optimizer: Adam and learning rate: 0.0001 by 96.42%. The model was tested using a testing dataset and obtain accuracy of 96% based on the confusion matrix. The findings indicate that combining CNN with fluorescence images improved the ability to classify the amount of aflatoxin contamination in cacao beans. This method has the potential to be more accurate and economical than the current approach, which could be adapted to reduce aflatoxin's negative effects on food safety and cacao trade losses.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"94 8","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140511064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This study explores the incorporation of time-varying sign gain into a parallel iterative learning control (ILC) architecture, augmented by an expert system, to enhance the performance and stability of a robotic arm system. The methodology involves iteratively tuning the learning control gains using time-varying sign gain guided by an expert system. Stability analysis, encompassing asymptotic and monotonic convergence, demonstrates promising results across multiple joints, affirming the effectiveness of the proposed control architecture. In comparison with traditional PID control, fixed gain ILC, and ILC with adaptive learning in the expert system, the analysis focuses on stability, precision, and adaptability, using root mean square error (RMSE) as a key metric. The results show that ILC with adaptive learning from the expert system consistently reduces RMSE, even in the presence of learning transients. This adaptability effectively controls the learning transients, ensuring improved performance in subsequent iterations. In conclusion, the integration of time-varying sign gain with expert system assistance in a parallel ILC architecture holds promise for advancing adaptive control in robotic systems. Positive outcomes in stability, precision, and adaptability suggest practical applications in real-world scenarios. This research provides valuable insights into the implementation of dynamic learning mechanisms for enhanced robotic system performance, laying the groundwork for future refinement in robotic manipulator control systems.
{"title":"Adaptive Parallel Iterative Learning Control with A Time-Varying Sign Gain Approach Empowered by Expert System","authors":"Phichitphon Chotikunnan, Rawiphon Chotikunnan, Panya Minyong","doi":"10.18196/jrc.v5i1.20890","DOIUrl":"https://doi.org/10.18196/jrc.v5i1.20890","url":null,"abstract":"This study explores the incorporation of time-varying sign gain into a parallel iterative learning control (ILC) architecture, augmented by an expert system, to enhance the performance and stability of a robotic arm system. The methodology involves iteratively tuning the learning control gains using time-varying sign gain guided by an expert system. Stability analysis, encompassing asymptotic and monotonic convergence, demonstrates promising results across multiple joints, affirming the effectiveness of the proposed control architecture. In comparison with traditional PID control, fixed gain ILC, and ILC with adaptive learning in the expert system, the analysis focuses on stability, precision, and adaptability, using root mean square error (RMSE) as a key metric. The results show that ILC with adaptive learning from the expert system consistently reduces RMSE, even in the presence of learning transients. This adaptability effectively controls the learning transients, ensuring improved performance in subsequent iterations. In conclusion, the integration of time-varying sign gain with expert system assistance in a parallel ILC architecture holds promise for advancing adaptive control in robotic systems. Positive outcomes in stability, precision, and adaptability suggest practical applications in real-world scenarios. This research provides valuable insights into the implementation of dynamic learning mechanisms for enhanced robotic system performance, laying the groundwork for future refinement in robotic manipulator control systems.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"25 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140513016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Dimas Chaerul Ekty Saputra, Alfian Ma'arif, K. Sunat
This study addresses the necessity for more advanced diagnostic tools in managing diabetes, a chronic metabolic disorder that leads to disruptions in glucose, lipid, and protein metabolism caused by insufficient insulin activity. The research investigates the innovative application of machine learning models, specifically Stacked Multi-Kernel Support Vector Machines Random Forest (SMKSVM-RF), to determine their effectiveness in identifying complex patterns in medical data. The innovative ensemble learning method SMKSVM-RF combines the strengths of Support Vector Machines (SVMs) and Random Forests (RFs) to leverage their diversity and complementary features. The SVM component implements multiple kernels to identify unique data patterns, while the RF component consists of an ensemble of decision trees to ensure reliable predictions. Integrating these models into a stacked architecture allows SMKSVM-RF to enhance the overall predictive performance for classification or regression tasks by optimizing their strengths. A significant finding of this study is the introduction of SMKSVM-RF, which displays an impressive 73.37% accuracy rate in the confusion matrix. Additionally, its recall is 71.62%, its precision is 70.13%, and it has a noteworthy F1-Score of 71.34%. This innovative technique shows potential for enhancing current methods and developing into an ideal healthcare system, signifying a noteworthy step forward in diabetes detection. The results emphasize the importance of sophisticated machine learning methods, highlighting how SMKSVM-RF can improve diagnostic precision and aid in the continual advancement of healthcare systems for more effective diabetes management.
{"title":"Optimizing Predictive Performance: Hyperparameter Tuning in Stacked Multi-Kernel Support Vector Machine Random Forest Models for Diabetes Identification","authors":"Dimas Chaerul Ekty Saputra, Alfian Ma'arif, K. Sunat","doi":"10.18196/jrc.v4i6.20898","DOIUrl":"https://doi.org/10.18196/jrc.v4i6.20898","url":null,"abstract":"This study addresses the necessity for more advanced diagnostic tools in managing diabetes, a chronic metabolic disorder that leads to disruptions in glucose, lipid, and protein metabolism caused by insufficient insulin activity. The research investigates the innovative application of machine learning models, specifically Stacked Multi-Kernel Support Vector Machines Random Forest (SMKSVM-RF), to determine their effectiveness in identifying complex patterns in medical data. The innovative ensemble learning method SMKSVM-RF combines the strengths of Support Vector Machines (SVMs) and Random Forests (RFs) to leverage their diversity and complementary features. The SVM component implements multiple kernels to identify unique data patterns, while the RF component consists of an ensemble of decision trees to ensure reliable predictions. Integrating these models into a stacked architecture allows SMKSVM-RF to enhance the overall predictive performance for classification or regression tasks by optimizing their strengths. A significant finding of this study is the introduction of SMKSVM-RF, which displays an impressive 73.37% accuracy rate in the confusion matrix. Additionally, its recall is 71.62%, its precision is 70.13%, and it has a noteworthy F1-Score of 71.34%. This innovative technique shows potential for enhancing current methods and developing into an ideal healthcare system, signifying a noteworthy step forward in diabetes detection. The results emphasize the importance of sophisticated machine learning methods, highlighting how SMKSVM-RF can improve diagnostic precision and aid in the continual advancement of healthcare systems for more effective diabetes management.","PeriodicalId":443428,"journal":{"name":"Journal of Robotics and Control (JRC)","volume":"60 6","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-01-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140513336","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：