IEEE Transactions on Automation Science and Engineering最新文献

{"title":"Adaptive Intrusion-Tolerant Control for Constrained MIMO Nonlinear Cyber-Physical Systems Against Hybrid Cyber Attacks","authors":"Zhaoyang Cuan;Yingying Ren;Da-Wei Ding","doi":"10.1109/TASE.2026.3667566","DOIUrl":"10.1109/TASE.2026.3667566","url":null,"abstract":"In this paper, we are concerned with the tracking control issue for a category of constrained MIMO nonlinear cyber-physical systems (CPSs) with unidentified disturbances, which face the exposure to venomous hybrid cyber attacks. Distinct from most existing related researches which can only address the intrusion-tolerant control issue against single type of cyber attacks, the studied problem in this paper that has not been well resolved is more challenging, where multiple types of network attacks including denial-of-service (DoS) attacks and deception attacks may coexist simultaneously, increasing the design difficulty of intrusion-tolerant controller. Furthermore, by means of designing two error transformation functions (ETFs) and proposing a novel error transformation design technique for resolving the nonlinearity and MIMO coupling problem on the basis of quantifying the destruction mechanism of cyber attacks, a novel adaptive tracking control scheme is ulteriorly developed. As a result, the proposed control architecture owns inherent intrusion-tolerant capability and robustness, which is capable of settling exogenous disturbances and hybrid cyber attacks in unison. In particular, the tracking errors can be forced to convert into the preassigned constrained region at pregiven decay rate in preseted settling time. Finally, the feasibility and superiority of the proposed control framework are illustrated through an emblematic application instance. Note to Practitioners—From the perspective of security defense, great progress has been made in the security research of constrained nonlinear CPSs, such as vulnerability analysis, attack detection and intrusion-tolerant control. However, once the defense mechanism preseted on the cyber layer crashes, the physical side may suffer cross-domain intrusion by malicious adversaries, resulting in the catastrophic non-contact damage. What is worse, the constrained nonlinear CPSs are vulnerable to the simultaneous existing hybrid cyber attacks and unidentified disturbances. Therefore, this paper investigates the tracking control issue for a category of constrained MIMO nonlinear CPSs with unidentified disturbances, which face the exposure to venomous hybrid cyber attacks. By means of designing two error transformation functions (ETFs) and proposing a novel error transformation design technique for resolving the nonlinearity and MIMO coupling problem on the basis of quantifying the destruction mechanism of cyber attacks, a novel adaptive intrusion-tolerant control scheme is ulteriorly developed. As a result, the proposed control architecture owns inherent intrusion-tolerant capability and robustness, which is capable of settling exogenous disturbances and hybrid cyber attacks concurrently. Notably, the tracking errors can be forced to convert into the preassigned constrained region at pregiven decay rate in preseted settling time.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5313-5322"},"PeriodicalIF":6.4,"publicationDate":"2026-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278546","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"AI-Driven Attention and Integer Programming for Optimal Motion Planning and Control in Autonomous Vehicles","authors":"Zhengtian Wu, Fanya Sun, Qing Gao, Jianbin Qiu","doi":"10.1109/tase.2026.3667574","DOIUrl":"https://doi.org/10.1109/tase.2026.3667574","url":null,"abstract":"","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"45 1","pages":""},"PeriodicalIF":5.6,"publicationDate":"2026-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147279997","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptive Fractional Observer-Based Frequency Regulation in Renewable-Integrated Power Systems","authors":"Abhishek Saxena;Ehab El-Saadany;Hatem H. Zeineldin;Sanjoy Kumar Parida","doi":"10.1109/TASE.2026.3666976","DOIUrl":"10.1109/TASE.2026.3666976","url":null,"abstract":"The penetration of Renewable Energy Resources (RERs) presents a significant challenge to the functionality and reliability of electrical power grid. Issues such as largescale RERs integration and load variability create substantial operational difficulties, hampering system inertia and affecting grid frequency control. Therefore, adaptive inertia management in the considerable renewable integrated grid is the main concern of this investigation. In order to enhance the frequency control in the test grid, a unified model free adaptive fractional state observer driven dynamic control methodology is developed for unpredictable dynamic conditions. The technique improves the system’s ability to reduce the oscillations and disruptions frequently experienced by RERs by permitting adaptive inertia. By enabling adaptive inertia, the strategy enhances the system’s ability to mitigate the oscillations and disruptions which are generally associated with RERs. A comparative evaluation with the robust existing control algorithms demonstrates the supremacy of the proposed control strategy. The findings are further validated through real-time experimental testing using OPAL-RT Hardware-In-the-Loop framework. Note to Practitioners—The rapid integration of RERs into present power systems poses critical challenges to instability under dynamic conditions. Therefore, this work presents a modelfree adaptive fractional state observer-based control strategy for practitioners involved in renewable-integrated power system design and control. This method offers a robust and flexible solution capable of handling system uncertainties without relying on system modeling. Further, a real time validations by experiments clearly shows its usefulness for improving modern grids stability issues due to massive RERs integration. In the future, the proposed methodology can be extended for real-world installation across intricate and changing grid circumstances by integrating EV fleet batteries for energy management and utilizing machine learning for further resilience.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5301-5312"},"PeriodicalIF":6.4,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Dynamic-Gain-Based Tunable Transient Performance for Unknown Multi-Agent Systems Under DoS Attacks and Disturbances","authors":"Yajing Zhao;Xiaoping Liu;Cungen Liu","doi":"10.1109/TASE.2026.3667129","DOIUrl":"10.1109/TASE.2026.3667129","url":null,"abstract":"A resilient control scheme with tunable transient performances is proposed for multi-agent systems subjected to DoS attacks and external disturbances. A novel sufficiently differentiable distributed filter is introduced to eliminate control input singularities caused by differentiating the reference output signal, with relaxed parameter limitations compared to existing filters. Then, a dynamic control gain strategy, consisting of a smooth gain elevation phase and an online gain adaptation phase, is developed to achieve tunable transient performance. The smooth gain elevation phase coordinates the settling time by initially assigning a low convergence rate and increasing it as the agents approach the desired states. Additionally, this property also suppresses the initial control input amplitude, thereby preventing agents from going out of control due to excessively high control inputs. The online gain adaptation phase enhances robustness against external disturbances by rapidly and smoothly increasing the control gain for affected agents. The dynamic gain is designed to be sufficiently differentiable to avoid singularities in control inputs resulting from differentiating virtual controllers. Finally, simulation results on both numerical and tower crane multi-agent systems, compared with three existing methods, validate the effectiveness of the proposed control scheme. Note to Practitioners—Multi-agent systems, such as autonomous vehicles, robotics, and smart grids, are vulnerable to cyberattacks and disturbances, which can destabilize conventional controllers. Traditional controllers usually select fixed parameters by trial and error and produce large initial control signals. This work proposes a resilient control scheme that applies a smooth and dynamic parameter coordination strategy to suppress excessively high control signals, adjust settling time, and adapt to cyberattacks and external disturbances. Simulation studies, including a numerical multi-agent system and a tower crane multi-agent system, show that this approach improves both stability and resilience compared with existing methods, offering a practical method for safer and more robust control in multi-agent systems.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5424-5435"},"PeriodicalIF":6.4,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improved Robust Model Predictive Current Control via High-Order Terminal Sliding Mode Control for Permanent Magnet Synchronous Motor","authors":"Yanshu Gong;Ximei Zhao;Lei Wang","doi":"10.1109/TASE.2026.3667162","DOIUrl":"10.1109/TASE.2026.3667162","url":null,"abstract":"The traditional model predictive current control (MPCC) is sensitive to parameter changes, and it is difficult to balance the dynamic and steady-state performance of the current. The PI controller used in the speed loop has poor robustness and tracking accuracy. Therefore, an improved robust MPCC with cascaded high-order fast terminal sliding mode speed controller (HOFTSMSC-IRMPCC) based on high-order sliding mode observer (HOSMO) is proposed. Firstly, a HOSMO is designed to estimate and compensate the parameter changes of the current loop. A dynamic strategy of switching different MPCC schemes according to the running state of motor is proposed, which takes into account the characteristics of fast response of single-vector MPCC and excellent steady-state performance of three-vector MPCC. Then, high-order fast terminal sliding mode speed controller (HOFTSMSC) is proposed to replace the speed loop PI controller. Different from the traditional high-order sliding mode control (HOSMC) scheme, the first-order integral fast sliding mode surface is nested in the second-order non-singular terminal sliding mode surface, which weakens system chattering and improves the convergence speed and tracking accuracy. To further improve the robustness of the speed loop, a HOSMO with the same structure as the current loop is designed. Experimental results demonstrate that HOFTSMSC-IRMPCC has excellent control performance. Note to Practitioners—PMSM is widely used in industrial automation, electric vehicles, aerospace and other fields, and is favored for its high efficiency and excellent dynamic response characteristics. However, the traditional MPCC is very sensitive to parameter changes, and it is difficult to balance the dynamic and steady-state performance of the current. In addition, the PI controller in the speed loop has poor robustness and low tracking accuracy. To solve these problems, an improved robust MPCC with cascaded high-order fast terminal sliding mode speed controller (HOFTSMSC-IRMPCC) is proposed in this paper. We carried out experiments on a semi-physical simulation platform to verify the superiority of the method.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5237-5249"},"PeriodicalIF":6.4,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Synchronized Coupled Iterative Learning Control Addressing the Nonlinear Disturbances in Gantry Platforms","authors":"Her-Terng Yau;Yu-Tsun Chen;Po-Lun Lin","doi":"10.1109/TASE.2026.3667105","DOIUrl":"10.1109/TASE.2026.3667105","url":null,"abstract":"The demand for high-precision and high-efficiency motion control in modern industrial applications such as laser processing, photolithography, and nanotechnology has highlighted critical challenges in multi-axis control systems, particularly regarding synchronization and coupling dynamics. Traditional single-axis control approaches often neglect the mechanical coupling effects that arise from system structure, thereby limiting overall performance. To address this limitation, this paper proposes a novel Synchronized Coupled Iterative Learning Controller (SCILC) for systems driven by iron-core linear motors that inherently suffer from nonlinear disturbances such as friction and cogging force. By incorporating a coupled system model and embedding synchronization error into the learning framework, SCILC effectively accounts for inter-axis dynamics and enhances control performance. Comparative experimental results demonstrate that the proposed SCILC significantly outperforms conventional multi-loop control and Coupled Control Iterative Learning Control (CCILC) methods. Specifically, the root mean square (RMS) tracking error is reduced from <inline-formula> <tex-math>$128.917~mu $ </tex-math></inline-formula>m (traditional control) to <inline-formula> <tex-math>$3.255~mu $ </tex-math></inline-formula>m (CCILC), and further down to <inline-formula> <tex-math>$1.641~mu $ </tex-math></inline-formula>m using SCILC. In terms of synchronization, the RMS error decreases from <inline-formula> <tex-math>$1.39~mu $ </tex-math></inline-formula>rad to <inline-formula> <tex-math>$0.98~mu $ </tex-math></inline-formula>rad, achieving a 28.7% improvement. These findings validate that SCILC offers superior convergence, precision tracking, and synchronization accuracy, thereby making it well-suited for high-end industrial applications. Note to Practitioners—In industrial multi-axis platforms, precise trajectory tracking and synchronization control are essential for ensuring product quality and production efficiency. However, traditional control methods typically do not consider coupling effects and treat them merely as disturbances. This study proposes a SCILC, which effectively addresses the challenges caused by friction, cogging force, and structural coupling through coupled system modeling and a synchronization error-guided learning mechanism, thereby enhancing synchronization during system operation. Experimental results demonstrate that SCILC achieves sub-micron-level tracking accuracy and microradian-level synchronization precision even under significant nonlinear disturbances. This approach is suitable for industrial applications that demand high control performance without substantially increasing system complexity or cost.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5289-5300"},"PeriodicalIF":6.4,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278621","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bandwidth-Efficient Collaborative SLAM: Mutual Information for Data Selection and Finite-State Entropy for Lossless Compression","authors":"Weiqiang Zhang;Lan Cheng;Jiangxia Wei;Xinying Xu;Zhe Zhang;Bin Xin","doi":"10.1109/TASE.2026.3666939","DOIUrl":"10.1109/TASE.2026.3666939","url":null,"abstract":"Collaborative SLAM systems allow resource-constrained robots to offload intensive computation to a central server, making real-time mapping and localization feasible. However, this approach introduces significant communication overhead, which can hinder practical deployment due to the high bandwidth required to transmit sensor data. To address this communication bottleneck in large-scale SLAM systems, we propose a novel bandwidth-optimization strategy that enables efficient collaboration between resource-constrained robots and a powerful central server. This strategy leverages mutual-information theory to select and transmit only the most informative keyframes, significantly reducing redundancy. Additionally, we apply finite-state entropy coding to achieve lossless compression of the selected data. Experiments on the public dataset and real-world scene demonstrate that our approach achieves a substantial 78% reduction in bandwidth usage compared to state-of-the-art centralized multi-robot SLAM methods, while maintaining high-precision localization and mapping performance without introducing significant computational overhead. Note to Practitioners—In real-world multi-robot SLAM systems especially for mapping tasks, limited communication bandwidth can be a serious bottleneck for central coordination. This paper introduces a bandwidth-efficient approach that allows robots to only send the most informative visual keyframes, minimizing communication overhead. By using mutual information for keyframe selection and applying lossless finite-state entropy coding, the system reduces redundancy while maintaining high accuracy in localization and mapping. This method can be integrated into existing SLAM frameworks, boosting scalability and cutting down on wireless load without sacrificing performance. Additionally, we will open-source the full code, making it easier for practitioners to deploy and use in real-world applications.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5401-5412"},"PeriodicalIF":6.4,"publicationDate":"2026-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147278534","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimal Non-Anticipative All-Scenario-Feasible Scheduling Strategy for Community Microgrid With Vehicle-to-Grid Energy Storages","authors":"Ye Tang;Qian Yuan;Panpan Yang;Xuekai Wang;Gang Wang;Maode Yan","doi":"10.1109/TASE.2026.3666562","DOIUrl":"10.1109/TASE.2026.3666562","url":null,"abstract":"The rapid proliferation of electric vehicles (EVs), empowered by vehicle-to-grid (V2G) technology, provides a novel and scalable form of flexible energy storage. Integrating V2G-enabled EVs into community microgrid (C-MG) presents a promising solution to mitigate supply-demand imbalances induced by renewable energy sources. However, the optimal scheduling of C-MG with integrated V2G energy storage remains significant challenges, primarily due to modeling complexities arising from large system scale and stochastic user behaviors, as well as the non-anticipativity and all-scenario-feasibility requirements for optimal decision-making. To address these challenges, this paper proposes a multi-stage robust optimal method to obtain the non-anticipative all-scenario-feasible scheduling decisions for C-MG with V2G energy storage. Specifically, an EV aggregation model that incorporates uncertainties in charging energy demand and large-scale heterogeneous charging behaviors is given first. Then an exact aggregated robust feasible region (ARFR) for scheduling decisions is developed, which guarantees non-anticipativity and all-scenario-feasibility. Based on this ARFR, the optimal charging/discharging profiles for the charging station are determined by solving an equivalent mixed integer linear programming (MILP) problem aimed at minimizing operational costs. Comparative case studies across different methods, EV types and scaling cases confirm that the proposed method enhances stable microgrid operation under various uncertainties. Note to Practitioners—The main purpose of this paper is to leverage large-scale V2G-enabled EVs as flexible energy storage to mitigate renewable energy volatility in C-MG. The main challenges are to establish the large-scale EVs aggregation model and guarantee the solution feasibility. To make the problem tractable with feasible solution, an efficient method is established based on the Minkowski sum and the feasible condition which allows the decision maker to obtain the economic scheduling decisions of the V2G energy storage more efficiently. For practitioners, the proposed approach provides a practical framework for incorporating large numbers of storage-like resources into system operations. In addition, the scheduling strategy can be readily adapted to other multi-stage robust optimization problems with similar constraint structures, providing engineers and system operators with a versatile tool for improving grid flexibility and reliability.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5387-5400"},"PeriodicalIF":6.4,"publicationDate":"2026-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146230636","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Incremental Q-Learning for Data-Driven Adaptive Critic Control of Wastewater Treatment Processes With State Constraints","authors":"Mingming Zhao;Ding Wang;Junfei Qiao;Jin Ren","doi":"10.1109/TASE.2026.3666246","DOIUrl":"10.1109/TASE.2026.3666246","url":null,"abstract":"The wastewater treatment process (WWTP) can effectively convert wastewater into reusable water, with dissolved oxygen (DO) concentration playing a crucial role in treatment efficiency. To achieve online control of DO concentration, a data-driven incremental &lt;inline-formula&gt; &lt;tex-math&gt;$Q$ &lt;/tex-math&gt;&lt;/inline-formula&gt;-learning (IQL) method with utility reshaping is proposed. The core idea is to integrate the incremental information of control inputs, the norm value of tracking errors, and a control barrier function into the utility function. First, the IQL method designs an online controller by utilizing tracking errors and the incremental information of control inputs, which reduces the approximation pressure on neural networks caused by excessively large control inputs. Second, the IQL leverages the norm value of tracking errors to influence the magnitude of the incremental control inputs, thereby enhancing the adaptive capability and online stability. Third, an adjustable control barrier function is employed to ensure that the tracking errors do not exceed the predefined constraints, thus meeting the specified tracking accuracy requirements. Fourth, a stability criterion related to learning rates is established to demonstrate the convergence of neural network weights. Finally, the superior control performance of the IQL method is validated via simulation results under two distinct weather conditions. Note to Practitioners—The complex biochemical reaction processes, unknown system parameters, and excessively large control inputs pose significant challenges to the design of controllers in WWTPs. Through online interaction, the &lt;inline-formula&gt; &lt;tex-math&gt;$Q$ &lt;/tex-math&gt;&lt;/inline-formula&gt;-learning method in reinforcement learning can leverage operational data to design adaptive controllers. However, traditional &lt;inline-formula&gt; &lt;tex-math&gt;$Q$ &lt;/tex-math&gt;&lt;/inline-formula&gt;-learning methods directly take large control inputs as the output of the action network and the input of the critic network, which often leads to the divergence of weights in both networks. Moreover, the weight matrix &lt;inline-formula&gt; &lt;tex-math&gt;$R$ &lt;/tex-math&gt;&lt;/inline-formula&gt; for the control inputs in the utility function of traditional &lt;inline-formula&gt; &lt;tex-math&gt;$Q$ &lt;/tex-math&gt;&lt;/inline-formula&gt;-learning is difficult to determine, requiring a large number of repetitive experiments. To overcome these challenges, the IQL method replaces the control input with its increment and uses the norm value of the tracking error to adaptively balance the magnitude of the increment. Additionally, to meet practical accuracy requirements, the IQL approach integrates an adjustable control barrier function, which can constrain the range of tracking errors. By using the proposed IQL mechanism, engineers can more easily create effective critic-learning-based optimal controllers for complex environments with external interference and unknown dynamics. Finally, experimental trials demonstrate its exceptio","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5869-5883"},"PeriodicalIF":6.4,"publicationDate":"2026-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146230637","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Automated Instance Segmentation Network for Overlapping Cells in Cleavage-Stage Embryos","authors":"Chen Sun;Hang Liu;Guanqiao Shan;An Hu;Haocong Song;Wenyuan Chen;Lei Chen;Shanshan Wang;Haixiang Sun;Yu Sun","doi":"10.1109/TASE.2026.3665874","DOIUrl":"10.1109/TASE.2026.3665874","url":null,"abstract":"Quantitative analysis of cleavage-stage embryos is a critical step for embryo evaluation in in vitro fertilization (IVF) treatment. Deep learning-based instance segmentation methods have emerged as a promising solution. However, existing methods, typically designed for opaque instances in nature scenes, face challenges when applied to segmenting overlapping cells in cleavage-stage embryos. The semi-transparency of the cells in cleavage-stage embryos leads to vague boundaries and complex overlaps between cells, making it difficult for existing models to segment these cells without merging or missing cells. In this work, we propose a novel instance segmentation network for segmenting overlapping, semi-transparent cleavage-stage embryo cells. Specifically, the network introduces two specialized branches that learn complementary representations: an overlap branch for addressing inter-instance overlaps and a boundary branch for dealing with intra-instance shapes. Then, an attention-based feature fusion module is designed to integrate learned overlap and boundary features to enhance the instance representation for the segmentation branch. Finally, an overlap-aware post-processing module leverages the outputs from the overlap and boundary branches to adaptively adjust suppression thresholds, ensuring the preservation of distinct yet highly overlapping cells. Experiments on two cleavage-stage embryo datasets demonstrate that our method outperformed state-of-the-art methods in both segmentation accuracy and recall, highlighting its potential for quantitative embryo evaluation in clinical IVF. Note to Practitioners—This work aims to automate the subjective and time-consuming process of manual morphology evaluation of cleavage-stage embryos, a critical step for selecting the most viable embryos in in vitro fertilization (IVF) treatment. Existing methods for segmenting embryo cells struggle with the semi-transparent and overlapping nature of embryo cells, leading to merging or missing results. To address these issues, this paper introduces a dual-branch architecture to explicitly model inter-instance overlaps and intra-instance shapes for overlapping cells. An attention-based feature fusion module is designed to integrate learned overlap and boundary features, providing enhanced features for the segmentation branch. An overlap-aware post-processing module is designed to refine predictions by adaptively preserving distinct yet overlapping cells. Experiment results suggest that the proposed method outperformed existing instance segmentation methods, particularly in high-overlap scenarios. The proposed method offers an accessible tool for automated embryo evaluation and has applicability to analyzing other semi-transparent cell types in medical imaging.","PeriodicalId":51060,"journal":{"name":"IEEE Transactions on Automation Science and Engineering","volume":"23 ","pages":"5225-5236"},"PeriodicalIF":6.4,"publicationDate":"2026-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146231074","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}