The Manned Submersible Vehicle (MSV) is a critical tool for combating submarine oil pipeline leaks. This paper adopts the core idea of a goal-oriented control system (GOCS). We propose a dual control for exploration and exploitation (DCEE) framework. The DCEE guides the submersible to autonomously collect rich measurement data, enabling accurate and efficient source-term estimation. For the first time, we developed a plume model adapted to submarine pollutant dispersion to guide the MSV equipped with the DCEE framework for source pinpointing. Under uncertainty of leakage parameters, a search strategy is designed to maximize the information obtained to detect leak sources. Furthermore, we employ a disturbance observer-based control (DOBC) algorithm in the low-level control to compensate for external disturbances. It ensures stable path tracking and yields a closed-loop autonomous search system for underwater operations. The simulation results show that our method can guide MSV to the leak source in complex and unknown submarine environments, which realizes rapid and accurate localization.

This article introduces a novel framework for achieving leader-steered (L-S) rigid formations within a multirobot vehicle system subject to nonholonomic constraints, while considering field-of-view (FOV) constraints. In contrast to the conventional separation-bearing leader-follower model, this framework incorporates a virtual leader model, established through topological and local agent connections. To achieve L-S rigid formations and address FOV constraints, a transformative approach is employed. In addition to forming L-S rigid formations, the framework ensures visibility maintenance between topologically connected vehicles using onboard cameras. This is achieved through the introduction of a continuous and continuously differentiable switching function, crucial in balancing visibility maintenance with formation adjustments, particularly when the global leader traverses trajectory segments with large curvature. To implement the framework, the distributed control protocol and the distributed observer are developed. Numerical simulations and real-world experiments demonstrate the framework's capability to achieve L-S rigid formations while accommodating FOV constraints, showcasing its practical utility and effectiveness in real-world applications.

Task scheduling in heterogeneous distributed systems is critical for industrial platforms, where decisions must be made under strict time constraints while resource states evolve dynamically. Existing approaches face significant limitations: classical heuristics yield suboptimal solutions; metaheuristics scale poorly; learning-based methods require extensive training with limited generalization. This article proposes a transformer-initialized dual-population evolution (TIDE), integrating three innovations: first, enhanced graph coloring preprocessing for enriched task representation, second, Transformer-based cross-modal attention for intelligent initialization of feasible solutions without offline pretraining, supported by an online adaptation mechanism, and third, asymmetric dual-population cooperative optimization with adaptive dimensionality reduction. Comprehensive experiments demonstrate that TIDE consistently outperforms state-of-the-art metaheuristics by 8%–13% in makespan while achieving an 80%–85% reduction in algorithm computing time compared to the metaheuristic average. On real scientific workflows, TIDE improves resource utilization by 4%–6% and maintains load balance above 94%, while maintaining response times within industrial deadlines. These results establish TIDE as a scalable solution for real-time scheduling in large-scale industrial systems.

This article addresses distributed constrained multiobjective resource allocation problems (DCMRAPs) in multiagent networks, where agents face multiple conflicting local objectives under local and global constraints. By reformulating DCMRAPs as single-objective weighted$L\!_p$problems, the proposed approach enables distributed solutions without relying on predefined weighting coefficients or centralized decision-making. Leveraging prescribed-time control and dynamic event-triggered mechanisms (ETMs), a novel distributed algorithm is proposed within a prescribed time through sampled communication. Using generalized time-based generators (TBGs), the algorithm provides more flexibility in optimizing solution accuracy and trajectory smoothness without the constraints of initial conditions. Novel dynamic ETMs, integrated with generalized TBGs, improve communication efficiency by adapting to local error metrics and network-based disagreements, while providing enhanced flexibility in balancing solution accuracy and communication frequency. The Zeno behavior is excluded. Validated by Lyapunov analysis and simulation experiments, our method demonstrates superior control performance and efficiency compared to existing methods, advancing distributed optimization (DO) across diverse applications.

Dual Control for Exploitation and Exploration (DCEE) shows promising performance by realizing optimal trade-off between exploitation and exploration under an unknown environment. However, it is computationally intensive and lacks rigorously established properties such as stability and convergence. This paper addresses these two issues by developing the Nesterov Accelerated Gradient Descent (NAGD) based DCEE, i.e. DCEE-NAGD, where the NAGD is applied to both the source term estimation and the path planning in the DCEE framework. It shows that DCEE-NAGD significantly reduces the search time by driving the search agent moving towards the estimated airborne source location (exploitation) and actively searching new data to reduce the current estimation uncertainty (exploration) with the help of NAGD. The convergence of both the source term estimation and the path planning of the DCEE-NAGD algorithm is rigorously established by applying the mean value theorem and mathematical transformation. More specifically, the convergence boundaries and the convergence rates of the source term estimation and the whole DCEE-NAGD algorithm are rigorously established. Both theoretic analysis and simulations confirm the proposed DCEE-NAGD algorithm significantly improves the performance so reduces the autonomous search time.

The quest for optimal operation in environments with unknowns and uncertainties is highly desirable but critically challenging across numerous fields. This paper develops a dual control framework for exploration and exploitation (DCEE) to solve an auto-optimization problem in such complex settings. In general, there is a fundamental conflict between tracking an unknown optimal operational condition and parameter identification. The DCEE framework stands out by eliminating the need for additional perturbation signals, a common requirement in existing adaptive control methods. Instead, it inherently incorporates an exploration mechanism, actively probing the uncertain environment to diminish belief uncertainty. An ensemble based multi-estimator approach is developed to learn the environmental parameters and in the meanwhile quantify the estimation uncertainty in real time. The control action is devised with dual effects, which not only minimizes the tracking error between the current state and the believed unknown optimal operational condition but also reduces belief uncertainty by proactively exploring the environment. Formal properties of the proposed DCEE framework like convergence are established. A numerical example is used to validate the effectiveness of the proposed DCEE. Simulation results for maximum power point tracking are provided to further demonstrate the potential of this new framework in real world applications.Note to Practitioners—In numerous engineering applications, it is highly desirable to operate a system to improve the efficiency, enhance performance or save energy. However, attaining this optimal control is a challenging task, due to the presence of unknown system and/or environment parameters. We develop a principled approach to balance between exploration and exploitation, involving active learning to estimate unknown parameters and tracking the optimal operational condition based on current estimation. This paper provides a unified framework to solve general auto-optimization control problems. The simulation results demonstrate that the proposed method outperforms existing methods in terms of efficiency and optimality for maximum power point tracking problem, and it can be readily implemented for many other engineering problems. Future research include generalizing the proposed method to nonlinear systems, as well as exploring novel applications to facilitate the widespread adoption of our method.

This article proposes a tracking model for non-holonomic constraint robots, enabling the realization of leadersteered rigid geometry formations. By employing cameras and local leader-based approaches, the challenge posed by traditional separation-bearing control methods, which are incapable of establishing rigid formation for both translational and rotational control, is resolved. In addition, to maintain the connectivity of the sensing topology, the field-of-view (FOV) constraints of the on-board cameras are integrated into the controller design. A conversion approach is used to translate the FOV constraints into a rigid geometry formation. Additionally, there is a trade-off between visibility constraints and the leader-steered rigid geometry formation, particularly when the trajectory of the global leader has significant curvature. To address this problem, a continuously smooth transition function is employed. Ultimately, a fixedtime distributed control protocol and distributed observers are developed to realize the formation framework. Experimental results demonstrate that the proposed control protocol effectively achieves rigid geometric formations and satisfies FOV constraints.

In this article, a multi-estimator based computationally efficient algorithm is developed for autonomous search in an unknown environment with an unknown source. Different from the existing approaches that require massive computational power to support nonlinear Bayesian estimation and complex decision-making process, an efficient cooperative active-learning-based dual control for exploration and exploitation (COAL-DCEE) is developed for source estimation and path planning. Multiple cooperative estimators are deployed for environment learning process, which is helpful to improving the search performance and robustness against noisy measurements. The number of estimators used in COAL-DCEE is much smaller than that of the particles required for Bayesian estimation in information-theoretic approaches. Consequently, the computational load is significantly reduced. As an important feature of this study, the convergence and performance of COAL-DCEE are established in relation to the characteristics of sensor noises and turbulence disturbances. Numerical and experimental studies have been carried out to verify the effectiveness of the proposed framework. Compared with the existing approaches, COAL-DCEE not only provides convergence guarantee but also yields comparable search performance using much less computational power.

Resorting to the principle of rigid body kinematics, a novel framework for a multirobot network is proposed to form and maintain an invariant rigid geometric shape. Unlike consensus-based formation, this approach can perform both translational and rotational movements of the formation geometry, ensuring that the entire formation motion remains consistent with the leader. To achieve the target formation shape and motion, a distributed control protocol for multiple Euler-Lagrange robotic vehicles subject to nonholonomic constraints is developed. The proposed protocol includes a novel prescribed performance control (PPC) algorithm that addresses the second-order dynamics of the robotic vehicles by employing a combination of nonsingular sliding manifold and adaptive law. Finally, the effectiveness of the proposed formation framework and control protocol is demonstrated through the numerical simulations and practical experiments with a team of four robotic vehicles.

This paper proposes an active information-directed reinforcement learning (AID-RL) framework for autonomous source seeking and estimation problem. Source seeking requires the search agent to move towards the true source, and source estimation demands the agent to maintain and update its knowledge regarding the source properties such as release rate and source position. These two objectives give rise to the newly developed framework, namely, dual control for exploration and exploitation. In this paper, the greedy RL forms an exploitation search strategy that navigates the agent to the source position, while the information-directed search commands the agent to explore most informative positions to reduce belief uncertainty. Extensive results are presented using a high-fidelity dataset for autonomous search, which validates the effectiveness of the proposed AID-RL and highlights the importance of active exploration in improving sampling efficiency and search performance.

With consideration of tire force saturation in vehicle motions, a novel path-following controller is developed for autonomous vehicles with unknown-bound disturbances and unknown actuator parameters. An adaptive sliding-mode fault-tolerant control (ASM-FTC) strategy is designed to stabilize the path-following errors without any information of disturbance boundaries, actuator fault boundaries and steering ratio from the steering wheel to the front wheels. By selecting the distance from the center of gravity to the center of percussion as the preview length, the effects of the lateral rear-tire force are decoupled and cancelled out, and then the preview error, which represents the path-following performance, can be only commanded by the front-tire force. To further address the issue of unknown tire-road friction limits, a modified ASM-FTC strategy is presented to improve the path-following performance as the lateral tire force is saturated. Simulation results show that the modified ASM-FTC controller demonstrates superior tracking performance over the normal ASM-FTC while the autonomous vehicle follows desired paths.

Abstract The task of reconstructing detailed 3D human body models from images is interesting but challenging in computer vision due to the high freedom of human bodies. This work proposes a coarse-to-fine method to reconstruct detailed 3D human body from multi-view images combining Voxel Super-Resolution (VSR) based on learning the implicit representation. Firstly, the coarse 3D models are estimated by learning an Pixel-aligned Implicit Function based on Multi-scale Features (MF-PIFu) which are extracted by multi-stage hourglass networks from the multi-view images. Then, taking the low resolution voxel grids which are generated by the coarse 3D models as input, the VSR is implemented by learning an implicit function through a multi-stage 3D convolutional neural network. Finally, the refined detailed 3D human body models can be produced by VSR which can preserve the details and reduce the false reconstruction of the coarse 3D models. Benefiting from the implicit representation, the training process in our method is memory efficient and the detailed 3D human body produced by our method from multi-view images is the continuous decision boundary with high-resolution geometry. In addition, the coarse-to-fine method based on MF-PIFu and VSR can remove false reconstructions and preserve the appearance details in the final reconstruction, simultaneously. In the experiments, our method quantitatively and qualitatively achieves the competitive 3D human body models from images with various poses and shapes on both the real and synthetic datasets.

In this paper, a distributed method is proposed for training multiple generative adversarial networks (GANs) with private data sets via a game-theoretic approach. To facilitate the requirement of privacy protection, distributed training algorithms offer a promising solution to learn global models without sample exchanges. Existing studies have mainly concentrated on training neural networks using pure cooperation strategies, which are not suitable for GANs. This paper develops a new framework for distributed GANs, where two groups of discriminators and generators are involved in a zero-sum game. Under connected graphs, such a framework is reformulated as a constrained minmax optimisation problem. Then, a fully distributed training algorithm is proposed without exchanging any private data samples. The convergence of the proposed algorithm is established via advanced consensus and optimisation techniques. Simulation studies are presented to validate the effectiveness of the proposed framework and algorithm.

In this article, distributed algorithms are developed to search the generalized Nash equilibrium (NE) with global constraints. Relations between the variational inequality and the NE are investigated via the Karush-Kuhn-Tucker (KKT) optimal conditions, which provide the underlying principle for developing the distributed algorithms. Two time-varying consensus schemes are proposed for each agent to estimate the actions of others, by which a distributed framework is established. The algorithm with fixed-gains is designed with certain system knowledge, while the adaptive algorithm is proposed to address the problem when the system parameters are not available. The asymptotic convergence to the NE is established through the Lyapunov theory and the consensus theory. The power control problem in a femtocell network is formulated as a Nash game and is solved by the proposed algorithms. The simulation results are provided to verify the effectiveness of theoretical development.

This article aims to realize a cooperative active power control of doubly fed induction generator (DFIG)-based wind farm (WF) to maximize the total reactive power capacity while maintaining the active power supply-and-demand balance. Difficulties lie in that the accurate PQ-curve expressions of wind turbines therein are unknown and nonuniform, thereby putting an obstacle to distributed optimization. To address the problem, PQ-curve inaccuracy caused by expression simplification is analyzed through the bridge of rotor current frame, rotor overspeeding control prioritized operation is recommended, and a surrogate-assisted distributed optimization (SADO) scheme is proposed from the WF perspective. The proposed method iteratively uses measured operating data to prompt a surrogate model to fit the accurate model, and then the optimal control action is guaranteed by online exploitation-and-exploration process with demonstrated availability through convergence analysis. Further, coordination with offline pretraining ensures that convergence can be obtained within shortened iteration steps. Case studies on 150-MW DFIG WF demonstrate the effectiveness of the proposed SADO scheme regarding shortening the iteration number, a full extraction on reactive power capacity and the better performance for voltage support.

In this article, distributed algorithms are proposed for training a group of neural networks with private data sets. Stochastic gradients are utilized in order to eliminate the requirement for true gradients. To obtain a universal model of the distributed neural networks trained using local data sets only, consensus tools are introduced to derive the model toward the optimum. Most of the existing works employ diminishing learning rates, which are often slow and impracticable for online learning, while constant learning rates are studied in some recent works, but the principle for choosing the rates is not well established. In this article, constant learning rates are adopted to empower the proposed algorithms with tracking ability. Under mild conditions, the convergence of the proposed algorithms is established by exploring the error dynamics of the connected agents, which provides an upper bound for selecting the constant learning rates. Performances of the proposed algorithms are analyzed with and without gradient noises, in the sense of mean square error (MSE). It is proved that the MSE converges with bounded errors determined by the gradient noises, and the MSE converges to zero if the gradient noises are absent. Simulation results are provided to validate the effectiveness of the proposed algorithms.

3D human pose and shape estimation plays a vital role in many computer vision applications. There are many deep learning based methods attempting to solve the problem only relying on single-view RGB images for training the network. However, since some public datasets are captured from multi-view cameras system, we propose a novel method to tackle the problem by putting optimization-based multi-view model-fitting into a regression-based learning loop from multi-view images. Firstly, a convolutional neural network (CNN) regresses the pose and shape of a parametric human body model (SMPL) from multi-view images. Then, utilizing the regressed pose and shape as initialization, we propose an improved multi-view optimization method based on the SMPLify method (MV-SMPLify) to fit the SMPL model to the multi-view images simultaneously. Subsequently, the optimized parameters can be adopted to supervise the training of the CNN model. This whole process forms a self-supervising framework which can combine the advantages of the CNN approach and the optimization- based approach through a collaborative process. In addition, the multi-view images can provide more comprehensive supervision for the training. Experiments on public datasets qualitatively and quantitatively demonstrate that our method outperforms previous approaches in a number of ways.

In this article, a distributed multiobjective optimization problem is formulated for the resource allocation of network-connected multiagent systems. The framework encompasses a group of distributed decision makers in the subagents, where each of them possesses a local preference index. Novel distributed algorithms are proposed to solve such a problem in a distributed manner. The weighted$L_{p} $preference index is utilized in each agent since it can provide a robust Pareto solution to the problem. By using distributed fixed-time optimization methods, the$L_{p} $preference index is constructed online without specifying the unknown parameters. Then, it is proved that the problem admits a unique Pareto solution. By exploiting consensus and gradient descent techniques, asymptotic convergence to the optimal solution is established via Lyapunov theories. Distinct from most of the current works, the proposed framework does not require any prior information in the formulation process, and private data can be well protected using this distributed approach. Numerical examples are included to validate the effectiveness of the proposed algorithms.

With the development of Augmented Reality (AR), egocentric action recognition (EAR) plays an important role in accurately understanding demands from the user. However, EAR is designed to help recognize human-machine interaction in single egocentric view, thus difficult to capture interactions between two face-to-face AR users. Paired egocentric interaction recognition (PEIR) is the task to collaboratively recognize the interactions between two persons with the videos in their corresponding views. Unfortunately, existing PEIR methods always directly use linear decision function to fuse the features extracted from two corresponding egocentric videos, which ignore the consistency of interaction in paired egocentric videos. The consistency of interactions in paired videos, and features extracted from them, are correlated to each other. On top of that, we propose to derive the relevance between two views using bilinear pooling, which captures the consistency of two views in feature-level. Specifically, each neuron in the feature maps from one view connects to the neurons from the other view, which enforces the compact consistency between two views and then all possible paired neurons are used for PEIR. To be efficient, we use compact bilinear pooling with Count Sketch to avoid directly computing outer product. Experimental results on the PEV dataset shows the superiority of the proposed methods on the task PEIR.

Reconstruction of 3D human bodies, from images, faces many challenges, due to it generally being an ill-posed problem. In this paper we present a method to reconstruct 3D human bodies from multi-view images, through learning an implicit function to represent 3D shape, based on multi-scale features extracted by multi-stage end-to-end neural networks. Our model consists of several end-to-end hourglass networks for extracting multi-scale features from multi-view images, and a fully connected network for implicit function classification from these features. Given a 3D point, it is projected to multi-view images and these images are fed into our model to extract multiscale features. The scales of features extracted by the hourglass networks decrease with the depth of our model, which represents the information from local to global scale. Then, the multi-scale features as well as the depth of the 3D point are combined to a new feature vector and the fully connected network classifies the feature vector, in order to predict if the point lies inside or outside of the 3D mesh. The advantage of our method is that we use both local and global features in the fully connected network and represent the 3D mesh by an implicit function, which is more memory-efficient. Experiments on public datasets demonstrate that our method surpasses previous approaches in terms of the accuracy of 3D reconstruction of human bodies from images.

By involving only subtle motions of body parts, video-based microaction recognition is a very important but challenging problem. Most existing action recognition methods are developed for general actions, and the current state-of-the-art methods usually largely rely on high-layer features learned from convolutional neural networks (CNNs). High-layer CNN features usually contain more semantic information but less detailed information. However, detailed information can be important for microactions due to the motion subtleness of such actions. In this paper, we propose to more effectively learn midlayer CNN features for enhancing microaction recognition. More specifically, we develop a new dual-branch network for microaction recognition: one branch uses the high-layer CNN features for classification, and the second branch further explores the midlayer CNN features for classification. In the second branch, we introduce a novel subtle motion detector consisting of three modules: 1) a discriminative spatial-temporal feature learning module, which further learns the subtle motion features corresponding to the discriminative spatial-temporal regions, 2) a parallel multiplier attention module, which further refines the features learned in channels and spatial-temporal domains, and 3) an activation fusion module, which fuses the max and average activations from midlayer CNN features for classification. In the experiments, we build a new microaction video dataset, where the micromotions of interest are mixed with other larger general motions such as walking. Comprehensive experimental results verify that the proposed method yields new state-of-the-art performance in two microaction video datasets, while its performance on two generalaction video datasets is also very promising.

Estimating the 3D model of the human body is needed for many applications. However, this is a challenging problem since the human body inherently has a high complexity due to self-occlusions and articulation. We present a method to reconstruct the 3D human body model from a single RGB-D image. 2D joint points are firstly predicted by a CNN-based model called convolutional pose machine, and the 3D joint points are calculated using the depth image. Then, we propose to utilize both 2D and 3D joint points, which provide more information, to fit a parametric body model (SMPL). This is implemented through minimizing an objective function, which measures the difference of the joint points between the observed model and the parametric model. The pose and shape parameters of the body are obtained through optimization and the final 3D model is estimated. The experiments on synthetic data and real data demonstrate that our method can estimate the 3D human body model correctly.

Image segmentation can be used in non-destructive testing, tracking and recognition. Level set method for image segmentation has poor performance on efficiency. In this paper, we propose to use max-flow algorithm to optimize a locally improved Chan-Vese model for image segmentation in the presence of intensity inhomogeneity. The energy function of local Chan-Vese model is introduced firstly. This model consists of global term, local term and penalty term and the local term contributes the segmentation for images with intensity inhomogeneity. Then, we convert this energy function to the frame of Graph Cut whose energy function can be efficiently minimized by max-flow algorithm. As a result, the process of optimization of local Chan-Vese model can be accelerated by using max-flow algorithm. The experiments demonstrate that the proposed method can achieve satisfactory segmentation for images with intensity inhomogeneity as well as very high efficiency.

This study proposes an improved level set method for segmenting images with intensity inhomogeneity. One of the improvements is to consider the difference between an original image and an estimated image without bias field in the image model. Apart from using this difference, Gaussian distribution with means and variance is utilised as the local intensity descriptor to map the original image into another domain so the object and the background can be better separated in the transformed domain. Then, an improved level set energy function that combines the image term, local variance, and the above difference is defined. The minimisation of the function can be processed by level set evolution. The proposed method is compared with existing methods, and experiments on both synthetic and real images demonstrate that authors’ method has superior performance.

We present a Chinese word segmentation model learned from punctuation marks which are perfect word delimiters. The learning is aided by a manually segmented corpus. Our method is considerably more effective than previous methods in unknown word recognition. This is a step toward addressing one of the toughest problems in Chinese word segmentation.