IEEE/CAA Journal of Automatica Sinica

Dynamic Centerline Path–Planning for the Autonomous Locomotion of Magnetic Milli/Microrobots Using Magnetic Trap-Based Navigation Systems

Won–Il Song, Armando Ramos–Sebastian, Sung Hoon Kim

, Available online , doi: 10.1109/JAS.2026.126533

Abstract(11) HTML (0) PDF(3)

Abstract:
Magnetic microrobots have shown strong potential for targeted therapies in vascular environments; however, complex vessel geometries, flow disturbances, and limited imaging feedback often require manual operation, increasing the risk of vessel damage. Moreover, the lack of high-rate 3D tracking hinders the implementation of fully autonomous control systems. To address these challenges, we propose a magnetic trap-based 2D autopilot framework for real-time autonomous navigation of magnetic robots under limited-feedback conditions. The framework integrates a Trapping Point Error Compensation (TPEC) strategy, which improves open-loop control accuracy by empirically mapping desired trajectory points to experimentally reachable robot states and their corresponding control inputs, thereby compensating for system imperfections without position feedback. It also incorporates the Skeletonization and Quasi-Euclidean Path-Planning (SQEP) algorithm, which performs clearance-aware path-planning on a skeletonized channel representation to generate safe, executable trajectories while accounting for the robot’s physical dimensions. Experimental results show that TPEC achieves locomotion speeds approximately four times higher than those of a conventional PID controller while maintaining comparable accuracy, substantially reducing dependence on high-rate feedback. Comparative evaluations demonstrate that SQEP consistently generates clearance-aware trajectories across representative channel geometries, achieving an average centerline-adherence accuracy of 96.8%, mean deviations below one pixel (~67 µm), and computation times under 300 ms. The integrated framework enables robust autonomous navigation through confined channels with abrupt geometric variations and fluid flow disturbances. These results demonstrate the potential of the proposed methodology for autonomous navigation in controlled 2D confined environments and provide a basis for future extensions toward vascular applications.

LID-Driver: Lightweight Interactive Diffusion-Based Lane Change Decision for Autonomous Driving

Ruizhe Wang, Zhenhua Pan, Yuanqing Xia, Jie Li

, Available online , doi: 10.1109/JAS.2026.125882

Abstract(19) HTML (0) PDF(11)

Abstract:

Contrastive-Enhanced Spatial-Temporal Normality Learning for Multivariate Time Series Anomaly Detection and Interpretation

Huangyining Gao, Ruyue Xin, Peng Chen, Xi Li, Ang Bian, Yunni Xia

, Available online , doi: 10.1109/JAS.2026.126194

Abstract(27) HTML (0) PDF(12)

Abstract:
Multivariate time series anomaly detection and interpretation are crucial across domains like finance and system monitoring, where deviations from normal patterns often indicate failures or emerging risks. Although normality learning in a self-supervised manner can effectively capture patterns from an unlabeled time series, current approaches typically focus solely on learning either global or local temporal features and fail to model the normality of a time series from a spatial perspective. Furthermore, a lack of robustness to noise and disturbances can compromise the detection accuracy of these methods in complex scenarios. To address these limitations, we propose CESTaL, a C ontrastive- E nhanced S patio- T emporal norm a lity L earning framework. CESTaL introduces three novel components: 1) A Temporal Order and Relation Learning module (TORL) for temporal pattern consistency through permutation alignment and ranking constraints; 2) A Dynamic Graph Topology Learning (DGTL) module that automatically discovers spatial dependencies via adaptive graph networks; 3) A Contrastive Learning-based Normality Enhancement (CLNE) module learning noise-robust and invariant representations through contrastive learning. Additionally, we identify root cause dimensions through dimension-level reconstruction analysis and cross-view relation divergence. Extensive experiments on six benchmark datasets demonstrate that CESTaL achieves state-of-the-art detection performance, outperforming 10 baseline methods across both point-based and robust range-based evaluation metrics, and also delivers superior results on multiple anomaly explanation metrics, showing strong interpretability.

Learnable Reeds-Shepp Cost Map-Guided Dual-Critic Reinforcement Learning for Generalizable Autonomous Parking

Shuaicong Yang, Mengying Ruan, Yi Yang, Ting Zhang, Mengyin Fu

, Available online , doi: 10.1109/JAS.2026.126437

Abstract(16) HTML (0) PDF(8)

Abstract:
Autonomous parking in diverse and constrained environments remains a key challenge for intelligent vehicles. Existing learning-based approaches often suffer from poor generalizability and rely heavily on handcrafted, distance-driven rewards, leading to inefficient and unnatural behaviors in complex spaces. This work proposes a Reeds-Shepp (RS) cost map-guided dual-critic reinforcement learning (RL) framework to address these issues. First, an attention-based learnable RS cost map is constructed to encode environment-dependent pose feasibility for motion planning. Second, to improve the effectiveness of the RL method, a dual-critic architecture is introduced, decoupling Euclidean distance-based rewards from heuristic-guided feasibility-aware evaluations. The distance-based reward facilitates efficient task completion, while the heuristic-guided reward incorporates global maneuverability cost, guiding the agent to select feasible yet potentially non-intuitive maneuvers. Furthermore, a curriculum learning strategy is employed to accelerate convergence and enhance parking precision, particularly in confined parallel slots. Extensive experiments across multiple parking slot types, including parallel, vertical, and slant, demonstrate that the proposed method achieves a 97.3% average success rate, outperforming the baselines by up to 24.5%. Notably, for the most challenging parallel parking scenario, our method achieves a 92.9% success rate, which still represents an 8.3% improvement over the best baseline. The results in both simulation and real-vehicle experiments highlight the superior generalizability, safety, and convergence efficiency of the proposed method.

Barrier Lyapunov-Based Adaptive Control of a Disturbed Rotating Flexible Beam System With Boundary Input Dead Zone

Renjing Zhang, Yong Ren, Yang Yu, Hai-Long Pei, Jinpeng Yu

, Available online , doi: 10.1109/JAS.2026.126221

Abstract(17) HTML (0) PDF(9)

Abstract:
In this study, we consider the adaptive boundary control problem of a rotating flexible beam system (RFBS) comprised of a cantilevered beam with a tip payload connected to its upper end. The opposite end of the beam is fixed to the centre of a rotating rigid disk. We suppose that the dynamic process of the system is affected by unknown disturbances and parameters. The external control action consists of a control torque applied to the disk and a control exerted on the tip payload. We further assume that both control inputs have dead zone nonlinearity. First, the dead zone nonlinearity is divided into a desired control signal and a nonlinear input error. Second, adaptive boundary control and parameter compensation laws are designed for the RFBS to ensure vibration reduction and regulate the rotating speed of the disk to a desired value, by summing those input errors and external disturbances. Third, a barrier Lyapunov function (BLF) is provided to handle boundary constraints, guaranteeing that the system states keep within a predefined limit. Furthermore, the constructed control schemes ensure ultimate uniformity and boundedness regulation of the state variables, which is proved through the Lyapunov direct method. Finally, the effectiveness and robustness of our designed controllers are verified via numerical simulations.

Pixel-by-Pixel Deblurring of An Optical Remote Sensing Satellite Image With Spatial-Variant Blur

Hanyu Hong, Shuai Guo, Liang Ye, Lei Wang, Bo Yang, Ying Zhu, Tianxu Zhang

, Available online , doi: 10.1109/JAS.2026.126026

Abstract(50) HTML (0) PDF(21)

Abstract:
When remote sensing satellites observe the Earth, the acquired images suffer from spatial-variant blurring degradation due to atmospheric turbulence and cross-medium factors. It means that the degree and causes of blurriness vary across different pixels. Traditional methods address such issues by applying a uniform blur kernel for restoration of the whole image or local blocks. They often fail to produce sufficiently clear results. To tackle this problem, this paper proposes a pixel-level deblurring method. Firstly, we combine gradient-domain structure computation with saliency detection algorithms. This allows us to extract multiple regions from the degraded image where valuable information is relatively concentrated. We calculate the point spread function (PSF) of these regions with a constraint by saliency values as blur kernels, and estimate the blur kernels at each pixel location based on spatial relationships. Finally, we solve for each pixel value in our proposed model rather than directly deconvolve with blur kernels, and refine the kernels and pixel values using iterative optimization. Experiments demonstrate that, as an unsupervised approach, our proposed method can achieve better performance and be more efficient than state-of-the-art methods, and even some deep-learning methods in real spatially variant blurred remote-sensing images.

A Novel Fractional-Order PID Autotuning Method With Robustness to Time Constant Variations

Marcian D. Mihai, Robin De Keyser, Isabela R. Birs, Clara M. Ionescu, Cristina I. Muresan

, Available online , doi: 10.1109/JAS.2026.126236

Abstract(90) HTML (0) PDF(18)

Abstract:
Fractional-order PID controllers have received considerable attention due to their enhanced flexibility and potential improvements in disturbance rejection, performance and robustness. This paper introduces a model-free fractional-order Proportional Integrative Derivative (FO-PID) autotuning framework that explicitly targets robustness to dominant time constant variations, a robustness aspect that is largely unaddressed by current fractional-order autotuners. The process is locally parametrised using a single sine test, which provides the frequency-response point and slope information required for tuning. Based on these data, the process is assigned to a low-order or high-order family, and a corresponding target open-loop slope $dL$ is selected from a nominal guideline table. The controller is then tuned by adjusting the fractional-orders (via Levenberg--Marquardt optimisation or a simple grid search) to match the target $dL$, while the remaining controller gains and time constants are computed directly from the tuning equations. Fractional operators are implemented through Oustaloup’s recursive approximation. Extensive numerical studies on representative process families show that the resulting FO-PID controllers preserve phase margin and overshoot over wide dominant time constant changes, thereby yielding near-invariant closed-loop behaviour. Overall, the main contribution is a practical, experimentally driven autotuning method that transfers time constant-robustness concepts into the fractional-order PID setting without requiring a parametric plant model, providing a basis for extensions toward delay-robust, adaptive, and multi-objective variants.

Cerebellum-Inspired Visual Servoing and Orientation Control for Redundant Manipulators

Duojicairang Ma, Renpeng Huang, Long Jin

, Available online , doi: 10.1109/JAS.2026.126302

Abstract(77) HTML (0) PDF(21)

Abstract:
Cerebellum-inspired control schemes simulate the motor control capabilities of the cerebellum and have achieved notable progress in trajectory tracking tasks without addressing the challenges of visual servoing of the redundant manipulator. In this paper, a cerebellum model is constructed according to similarities between the cerebellum and the echo state network (ESN), and a cerebellum-inspired visual servoing and orientation control (CIVSOC) scheme is proposed based on the cerebellum model. The proposed scheme achieves the visual servoing control of the redundant manipulator while maintaining the stable orientation of the end-effector. Furthermore, to reduce the effect of noise, a Kalman filter (KF) is integrated into the control scheme and improves the robustness. Finally, simulations and experiments validate the effectiveness of the CIVSOC scheme in achieving both visual servoing and orientation control of the manipulator.

Safe Formation Tracking Control for Multi-Agent Euler-Lagrange Systems in Environments With Static and Dynamic Obstacles

Wei Cui, Yikang Yang, Yuxuan Guo, Ye Cao, Yongduan Song, Frank L. Lewis

, Available online , doi: 10.1109/JAS.2026.126038

Abstract(98) HTML (0) PDF(32)

Abstract:
Existing studies primarily address obstacle avoidance at the trajectory planning level and often neglect the subsequent tracking stage. However, inevitable tracking errors can still cause collisions even when the planned trajectory is deemed safe. This paper develops a safe formation-tracking control scheme for multi-agent Euler-Lagrange (EL) systems operating among multiple obstacles, resolving conflicts between formation tracking and collision avoidance. The obstacles considered include other agents, borders, enclosed static obstacles, and moving obstacles with unknown velocities. We propose an adaptive, distributed method that ensures collision avoidance while achieving asymptotic formation tracking. The key features are: i) it requires neither pre-acquired obstacle maps nor precise obstacle-velocity information; ii) unsafe reference trajectories are modified online based on locally sensed data to reconcile tracking and safety; and iii) integral barrier functions enter as multiplicative gains in the virtual control law, together with adaptive estimation of upper bounds on velocity-related terms, yielding safe formation tracking.

Robust Information Feature Pyramid Networks for Low-Visible Object Detection

Jie Hua, Zhongyuan Wang, Xin Tian, Jinsheng Xiao, Gang Wu, Tao Lu, Jiayi Ma

, Available online , doi: 10.1109/JAS.2026.126065

Abstract(115) HTML (0) PDF(21)

Abstract:
Detecting low-visible objects with missing contours, such as dark and foggy objects, presents challenges in computer vision. Previous methods have enhanced feature representation by training on synthetic datasets to improve image quality. However, the heavy reliance on paired and high-quality datasets renders data acquisition prohibitively expensive. In this work, we focus on low-quality images and propose a novel plug-and-play feature pyramid network named RIFPN, which can extract robust information. We first propose a novel depth-first search convolution (DFSconv) that progressively predicts regions containing effective features, enabling the capture of low-visibility object features across distant, cross-regional areas. We subsequently propose the Feature Integration Module (FIM), which projects multi-scale features to the same resolution and enables cross-level feature integration through DFSconv, achieving efficient interaction within the feature pyramid. Finally, we introduce the Feature Distribution Module (FDM), which has a varying feature search radius and adapts its search range based on feature resolution. This supplements low-visibility objects and provides more robust multi-scale features. Extensive experiments demonstrate that the robust information extracted using RIFPN significantly improves the detection accuracy of low-visible objects, including dark object detection (+2.2 AP on ExDark and +1.3 AP on DarkFace) and foggy object detection (+1.9 mAP on Foggy Cityscapes and 0.9 AP on HazyDet). In addition, our RIFPN, as a feature pyramid structure, improves 2.3 and 1.7 APs in the general object detection and instance segmentation tasks, respectively.

Distributed Cooperative Control for Common Object Manipulation With High Precision and Fast Convergence

Xingqiang Zhao, Yongduan Song

, Available online , doi: 10.1109/JAS.2026.126362

Abstract(161) HTML (0) PDF(41)

Abstract:
Coordinated object manipulation with strict demands on convergence precision and speed remains a major challenge in networked robotic systems. This paper develops a distributed control framework for networked robotic arms to cooperatively manipulate a common object. To achieve the cooperative task distribution in a distributed manner, a prescribed-time estimator (PTE) is developed for each arm to estimate the desired trajectory of the object. A significant merit of the proposed estimator is that zero-error estimation is achieved within the prescribed finite time, with the numerical singularity issue caused by infinite time-varying gains avoided. Together with the estimated task, a novel control policy is synthesized, where the pre-specified convergence precision is ensured within a prescribed time, including the following critical attributes: 1) the settling time and convergence precision are not influenced by the system’s initial conditions, order, or control parameters; and 2) the method is valid for an infinite time interval and allows the system’s initial states to be unknown or random. The effectiveness of the proposed strategy is validated through simulations.

Undersampled Signal Recovery With Uncertainty Reduction via Non-Uniform Sampling Strategy

Huahui Hu, Huan Liu, Laihao Yang, Zhibo Yang, Ruochen Jin, Konstantinos Gryllias

, Available online , doi: 10.1109/JAS.2026.126062

Abstract(171) HTML (0) PDF(26)

Abstract:
Blade tip timing (BTT) is a promising non-destructive testing technology for online monitoring of rotating blades. Due to physical constraints and measurement principles, BTT signals are highly under-sampled and accompanied by uncertainty and non-uniformity, which complicates subsequent signal processing. This paper reveals the inherent relationship between under-sampling, uncertainty, and non-uniformity. It innovatively transforms the aliasing frequency recovery problem into an uncertainty reduction problem. Undersampling introduces aliasing in the frequency domain, which means the uncertainty of signal recovery increases. De-aliasing is the process of uncertainty reduction, and the non-uniformity is utilized to achieve this goal. Based on this transformed problem, we propose a non-uniform sampling-based uncertainty reduction method (NUS-URM) to perform de-aliasing. Under the framework of Bayesian inference, a non-uniform sampling strategy using two BTT sensors is applied to recover the blade’s natural frequency from the aliased spectrum due to undersampling. Numerical and experimental validations demonstrate the feasibility and accuracy of the NUS-URM for signal recovery under weak prior conditions. The layout of the sensors and prior information are also discussed. After further analyzing the revealed relationships, the NUS-URM has the potential to achieve BTT signal recovery using a single sensor. Additionally, it may also be applied in fields such as computational photography, data compression, astronomical observation, and biomedical signal processing.

A Privacy-Preserving Distributed Optimization Algorithm With Prescribed-Time Convergence Using a Deep Learning Adversarial Network for an Integrated Energy System

Jianfei Chen, Ke Li, Zheng Qin, Daduan Zhao, Chenghui Zhang

, Available online , doi: 10.1109/JAS.2026.126008

Abstract(141) HTML (0) PDF(39)

Abstract:
This paper investigates the economic dispatch problem in multi-agent integrated energy systems (IES). To satisfy the stringent requirements for privacy preservation and convergence speed, a prescribed-time distributed optimization algorithm featuring an encryption mechanism based on a deep adversarial network is proposed. First, an economic dispatch model for an IES with coupled electrical and heating systems is established and formulated as a constrained distributed optimization problem. Second, a projection-based prescribed-time distributed algorithm is developed, and its convergence within a prescribed time is rigorously proven using Lyapunov theory. Subsequently, a deep learning-based encryption method, inspired by multi-agent game theory, is designed to enable rapid encryption and decryption of exchanged information while significantly enhancing data fidelity and security against attacks. Finally, case studies conducted on a modified IEEE 30-bus and 14-node test system, representing a coupled electrical and heating IES, demonstrate that the proposed method exhibits excellent convergence performance and robust privacy-preserving capabilities.

Enclosing Formation Control of Multi-Agent Systems under Denial-of-Service Attacks

Mahdi Baradarannia, Farzad Hashemzadeh, Tufan Kumbasar

, Available online , doi: 10.1109/JAS.2026.126161

Abstract(371) HTML (0) PDF(77)

Abstract:
Multi-Agent System (MAS) operation relies on networked communication, making it vulnerable to cyber-attacks. While much research has focused on how MASs reach consensus when they are under cyber threats, this paper is the first to address the enclosing formation problem with the same geometry as targets for a peer-to-peer communicating MAS in the presence of Denial-of-Service (DoS) attacks. First, a distributed practical finite-time estimator is proposed that enables followers to estimate the geometric center of target agents despite communication disruptions due to DoS attacks. Using Lyapunov functional analysis, it has been shown that the estimation error converges to a bounded region within a finite time. Building on this estimator, a control strategy is developed that ensures the system remains input-to-state stable even in the presence of DoS attacks. To facilitate control gain tuning, a parameter is introduced, providing a systematic approach for determining the controller gains. The approach is supported by rigorous theoretical analysis. To validate its effectiveness, a multi-agent surveillance scenario is simulated in which the communications between followers and targets experience aperiodic disruptions due to DoS attacks. The system shows robust performance during both attack and normal intervals, demonstrating its effectiveness in maintaining an enclosing formation under adversarial conditions.

Optimality-Guaranteed Model-Free Output Feedback Control Using Reinforcement Learning

Li Zhang, Jialu Fan, Tianyou Chai, Frank L. Lewis

, Available online , doi: 10.1109/JAS.2026.126134

Abstract(334) HTML (0) PDF(98)

Abstract:
This paper investigates a data-driven, model-free optimal output feedback (OPFB) control approach for linear time-invariant (LTI) continuous-time (CT) systems. Existing data-driven OPFB control methods are hindered by the challenge that the optimality of the control policy cannot be guaranteed within the OPFB framework. To overcome this limitation, the paper introduces two novel algorithms aimed at ensuring the optimal OPFB control solution. First, by redefining the traditional performance index, an offline learning algorithm based on an augmented Hamiltonian function (AHF) is introduced. This algorithm ensures the optimality of the OPFB control solution while eliminating the need for an initial control gain. Second, to extend the algorithm’s applicability to systems with unknown dynamics, a data-driven, model-free approach leveraging off-policy integral reinforcement learning (IRL) is developed. This method operates solely on system data, demonstrating robustness against probing noise and ensuring unbiased learning outcomes. The proposed algorithms are validated through numerical simulation experiments using an F-16 aircraft system, confirming their effectiveness in practical scenarios.

Towards the Bi-Imbalance: A Pruning-Based Contrastive Domain Adaptation Framework for Industrial Fault Diagnosis

Gecheng Chen, Zeyu Yang, Jianqiang Li, Xinkai Chen, Chengwen Luo

, Available online , doi: 10.1109/JAS.2026.126002

Abstract(472) HTML (0) PDF(57)

Abstract:
Industrial fault diagnosis often encounters both distributional shift and severe class imbalance across domains, termed bi-imbalance. Existing unsupervised domain adaptation (UDA) methods largely overlook this, leading to poor generalization or negative transfer. We propose a Pruning-based Contrastive Domain Adaptation (PCDA) framework to address bi-imbalance without relying on pseudo labels or manual weighting. PCDA integrates two key components: 1) An imbalance-aware contrastive learning (Ia-CLR) module that leverages model pruning to detect and emphasize minority samples during self-supervised pretraining; 2) A boundary-aware adversarial domain adaptation (Ba-ADA) module that enhances feature discriminability via pruned supervised contrastive learning. The pruning-induced perturbation implicitly re-weights hard-to-learn samples from both domains. Extensive experiments on benchmark datasets demonstrate that PCDA significantly outperforms state-of-the-art UDA methods under bi-imbalanced conditions.

Attribute-Augmented PPR Meets Self-Loops: Simple Yet Effective Defending Graph Neural Networks

Zeli Wang, Jie Li, Zhiqiu Ye, Longlong Lin, Mengdi Wang, Guoyin Wang

, Available online , doi: 10.1109/JAS.2026.125990

Abstract(575) HTML (0) PDF(57)

Abstract:
Graph Neural Networks (GNNs) are effective in processing graph-structured data but are also known to be vulnerable to adversarial attacks. Limitations of existing defense methodologies include a suboptimal approach to edge and node weight assignment, as such methods generally focus on the graph’s topological structure or node attributes. This imbalance can lead to inaccurate assessments of the importance of neighboring nodes, especially those targeted by attacks. We posit that the model should diminish the influence of insignificant and contaminated neighboring nodes on graph representation learning, thereby mitigating the adverse effects of compromised neighbors. Hence, we propose a versatile GNN adversarial defense framework by combining a novel attribute-augmented Personalized PageRank (PPR) with an adaptive self-loop weight adjustment mechanism, coined GPROP, to defend against adversarial attacks. Specifically, the attribute-augmented PPR helps compute the node similarity more accurately based on global semantics, by considering multi-hop graph topology and node attributes. According to the similarity scores, GPROP assigns corresponding weights to edges and simultaneously prunes insignificant ones. When iteratively aggregating the neighbor information, the proposed self-loop mechanism dynamically adjusts the weight ratio of neighboring nodes and the nodes themselves based on the node’s degree. This strategy reduces the influence of neighboring nodes that may introduce unreliable or harmful information. Moreover, GPROP is model-agnostic and can be readily embedded into different GNN backbones, leading to a substantial improvement in resistance to adversarial perturbations. Experimental results across multiple benchmark datasets show that our proposal consistently outperforms existing approaches, providing enhanced resilience and accuracy.

Reinforcement Learning-Based Adaptive Optimal Control for a Snake Robot

Yang Xiu, Zhiyi Shi, Guanghong Liu, Rob Law, Dongfang Li, Aiguo Song, Edmond Q. Wu

, Available online , doi: 10.1109/JAS.2025.125762

Abstract(1133) HTML (0) PDF(81)

Abstract:
Due to the difficulty of accurately modeling snake robots, model-based control schemes are ineffective, and the constraints of motion velocity and energy consumption pose challenges to meandering gait. In this work, a two-layer reinforcement learning-based adaptive optimal control framework for snake robots is proposed to achieve trajectory tracking motion of optimal energy efficiency gait. A multi-objective problem for gait amplitude, frequency, and phase is established in the optimization layer, which balances minimizing energy consumption and maximizing velocity by weighted summation. Multiple matching results of gait parameters and performance are obtained through proximal policy optimization, allowing users to select the optimal combination. In the control layer, an actor-critic-identifier neural network-based reinforcement learning optimal controller is designed by considering the difficulty in solving dynamics unknowns and Bellman equation. It adaptively fits the cost function and control policy, reducing the dependence on an accurate model and avoiding computational complexity. Theoretical analysis demonstrates that the proposed method can guarantee stability of tracking errors for snake robots, with optimal cost. Comparative simulation experiment results show the effectiveness and superiority of this method.

Majorization-Minimization-Based Neural Dynamics for Time-Variant Optimization Under Multi-Set Constraints

Ying Liufu, Yongji Guan

, Available online , doi: 10.1109/JAS.2026.125768

Abstract(1145) HTML (0) PDF(30)

Abstract:

Efficient Centralized Traffic Grid Signal Control Based on Meta-Reinforcement Learning

Jia Wu, Yican Lou

, Available online , doi: 10.1109/JAS.2023.123270

Abstract(2050) HTML (0) PDF(96)

Abstract:

Supplementary File of “Push-Sum Based Algorithm for Constrained Convex Optimization Problem and Its Potential Application in Smart Grid”

Qian Xu, Zao Fu, Bo Zou, Hongzhe Liu, Lei Wang

, Available online

Abstract(1492) HTML (0) PDF(1)

Abstract:

Supplementary Material for “Collision and Deadlock Avoidance in Multi-Robot Systems Based on Glued Nodes”

Zichao Xing, Xinyu Chen, Xingkai Wang, Weimin Wu, Ruifen Hu

, Available online

Abstract(2016) HTML (0) PDF(6)

Abstract:

Early Access

Links
More

E-mail Alert

Early Access

Links More

E-mail Alert

Links
More