This paper addresses the problem of safety-critical path-guided herding control for multiple-input multiple-output multi-agent systems under incomplete state measurement, safety constraints, and limited communication resources. Specifically, an output-feedback finite-time neural predictor is proposed to identify both model uncertainties and unknown state information. Subsequently, a distributed path-guided herding control strategy is designed, including patrolling, gathering, enclosing, and expelling. Control barrier functions are formulated as safety constraints, and a quadratic optimization problem is established. By using the neurodynamic optimization to solve the optimization problem, the optimal control law satisfying both safety constraints and state constraints is generated. Furthermore, a dynamic event-triggered communication method is proposed to reduce unnecessary communication, especially during the transient phase. By using the proposed herding control approach, the collision-free herding is guaranteed for input-to-state stability, and simulation results are provided to demonstrate the effectiveness of the approach.

This article investigates the pursuit–evasion problem of multiple underactuated autonomous surface vehicles (ASVs) under velocity and collision avoidance constraints. A safety–critical pursuit–evasion game (PEG) method based on min–max optimization and neural network dynamic control is proposed. Specifically, an allocation strategy is designed at first based on the position information of the pursuing and evading ASVs to achieve a rational and efficient allocation of pursuit targets by minimizing pursuit distances. Next, a nominal PEG guidance law is proposed by combining model predictive control (MPC) with min–max optimization methods. Then, the nominal guidance law is optimized based on a heading-constrained control barrier function (CBF) such that a safety–critical guidance law for collision avoidance can be achieved. Finally, a predicator-based neural network is developed to estimate the uncertainty and external disturbance, and a dynamic control law is proposed to track the guidance signals without using any model parameters. It is proven that the closed-loop system is input–to–state stable (ISS), and the ASV system is safe. A robot-operating-system (ROS)-based simulation results demonstrate the effectiveness of the proposed safety–critical PEG method based on min–max optimization and neural network dynamic control.

This paper investigates the safety-certified cooperative path following problem for second-order nonlinear systems with multiple-input multiple-output strict-feedback form subject to safety, state, and input constraints. A safety-certified learning control approach is proposed to achieve collision-free cooperative path following based on command optimization, online learning, and self-triggered communication. Specifically, an extended-state-observer-aided learning neural predictor is developed to simultaneously identify nonlinear functions and unknown input gains without measuring state derivatives. Control barrier functions are then employed to ensure safety through forward invariant sets. Next, command optimization is utilized to generate the optimal virtual control signals that satisfy safety constraints, state constraints, and input constraints. A neurodynamic optimization technique is employed to solve the quadratic optimization problem in real time. Additionally, a self-triggered mechanism is introduced in path variable coordination to reduce the listening and triggering times. By using the proposed safety-certified cooperative path following approach, a safe formation is guaranteed for input-to-state safety. Simulation results are provided to illustrate the effectiveness of the proposed method.Note to Practitioners—This paper addresses the safety-certified cooperative path following problem of multi-agent systems, which has practical implications in various applications. These applications include formation patrol, cargo transportation, search and rescue missions, swarm robotics, and agricultural tasks. By coordinating the movements of multiple agents, cooperative path following enhances efficiency in these scenarios. The challenges posed by safety, state, and input constraints are commonly encountered in practical. To address these challenges, the command optimization approach is proposed in this paper. The optimization problem is efficiently solved in real-time using neurodynamic optimization technique. Furthermore, to enhance feasibility in a limited communication environment, a self-triggered mechanism is introduced to reduce the communication burden. Therefore, the aforementioned effective scheme is suitable for implementation in industrial applications.

This paper focuses on a distributed cooperative guidance model-free control method for a cluster of under-actuated disk-type autonomous underwater gliders (AUGs) in the presence of unknown kinetic model parameters and ocean disturbances. Firstly, a distributed cooperative motion generator is designed to generate reference path points, and then a cooperative control method is proposed based on the update path parameters. Secondly, a three-dimensional (3D) guidance law is constructed by employing closed 3D vector fields. Finally, data-driven filtered adaptive extended state observers (DFAO) are proposed to deal with the unknown input gains, internal uncertainties and external disturbances of the disk-type AUGs, and an adaptive kinetic control law is designed by using the knowledge learned from the observers. Simulation results demonstrate the effectiveness of the proposed 3D distributed cooperative guidance model-free control method for disk-type AUGs subject to fully unknown kinetics. Note to Practitioners—The disk-type AUG has the characteristic of long operational endurance, capable of functioning continuously for several months when fully loaded. Consequently, this type of glider offers an advantage in establishing oceanic sensor networks. To achieve this goal, two technical challenges arise: the coordination among multiple gliders and the anti-disturbance control of individual glider. Our research focuses on distributed cooperative guidance and model-free control issues for multiple underwater gliders. First, we propose a distributed cooperative guidance scheme to maintain a specific formation among the gliders. Additionally, while ensuring control effectiveness, we employ data-driven methods to estimate uncertain kinetic model parameters. Our approach is not only theoretically viable but also ready for industrial application, thus filling a gap in underwater glider technology.

This article addresses the coordinated path-following (CPF) control under human supervision for marine vehicles (MVs) with unknown disturbances. A human-in-the-loop coordinated path-following (HCPF) control architecture is proposed based on the robust exact differentiator (RED) observer and the output feedback continuous twisting control (OFCTC) method. Specifically, a human manipulation is introduced into a virtual leader to regulate its path update speed in response to sudden circumstances for the follower MVs. All follower MVs synchronize indirectly with the human-in-the-loop virtual leader over a communication graph. Then, the path-following control problem is transformed into the controls of a second-order along-track error dynamic and a third-order cross-track error dynamic. By using the path-following errors only, the RED observers are developed to recover the unknown states of the error dynamics. Finally, the OFCTC laws are designed to achieve the individual path following in finite time regardless of the lumped disturbances. The stability of the closed-loop system is analyzed by employing the cascade theory. A salient feature of the proposed HCPF control architecture is that the CPF for MVs can be implemented under the supervision and intervention of human. Simulation results are given to illustrate the effectiveness of the proposed HCPF control method.

Recently, there has been significant progress in automatic pain assessment based on facial expression analysis. However, the performance of pain assessment remains unsatisfactory, due to a lack of analysis on local pain-related action units and emotional ambiguity. In particular, ambiguous pain expressions complicate the estimation of pain. It is argued that certain facial local regions related to pain should receive more attention while estimating pain intensities. Based on this, we propose a multi-task hybrid Conv-Transformer method for facial pain assessment, which utilizes the self-attention mechanism to explore facial local features related to pain intensities and constructs a multi-task joint optimizing module to mitigate facial emotional ambiguity. In particular, the proposed method modifies the network structure of the vision transformer model to better estimate continuous pain intensities. Meanwhile, a multi-task module is constructed to jointly optimize the classification and the regression tasks of pain assessment, which effectively regularizes the extracted features and facilitates a better fit of the regressed prediction to the given label. Finally, experimental results on the UNBC Pain dataset illustrate that the proposed method performs better with pain assessment compared with state-of-the-art methods.

In this paper, a finite set model predictive control method based on data-driven neural network predictors (DNNPs) is proposed for pulse width modulation (PWM) rectifiers with fully unknown parameters. First, DNNPs are structured based on concurrent learning such that model uncertainties and input gains are identified simultaneously. Secondly, based on the information estimated by the predictors, a finite set model predictive power controller is designed, which is responsible for simplifying the rolling optimization and reducing the computational complexity. Finally, the stability analysis is provided based on input-to-state stability theory, and simulation results are provided to prove the effectiveness of the proposed method.

This article investigates the yaw angle tracking control of an autonomous surface vehicle (ASV) subject to fully unknown internal dynamic, external disturbance, and unknown control input gain. A model-free adaptive antidisturbance autopilot control method is proposed for an ASV without using any model parameters. Specifically, by utilizing real-time and historical data, a data-driven concurrent learning extended state observer (CLESO) method is designed to estimate the unknown ASV model parameters and ensure the convergence of the estimation without requiring persistent excitation. Then, a model-free yaw angle tracking controller is designed based on the data-driven CLESO method. Through Lyapunov stability analysis, the closed-loop system is proven to be input-to-state stable. Simulation and experimental results validate the effectiveness of the proposed CLESO method for the yaw angle tracking of an ASV with fully unknown dynamic model.

This paper presents a finite control set model free predictive control method for doubly fed induction generator for direct current generation system. The system model is fully unknown, and only the input and output data are measured. Firstly, data-driven neural network predictors are proposed to reconstruct the system model. Filtered data of system states are recorded and used to update the neural network parameters. The system of the data-driven NN predictors is proved to be input-to-state stable. Then, a finite control set model free predictive controller is designed based on the perdiction model. The optimal control actions are selected based on the receding horizon optimization with a finite control set. Simulation results show the effectiveness of the proposed model free predictive control of doubly fed induction generator based on data-driven neural network predictors.

This article addresses the safety-critical containment maneuvering of multiple underactuated autonomous surface vehicles (ASVs) in the presence of multiple stationary/moving obstacles. In a complex marine environment, every ASV suffers from model uncertainties, external disturbances, and input constraints. A safety-critical control method is proposed for achieving a collision-free containment formation. Specifically, a fixed-time extended state observer is employed for estimating the model uncertainties and external disturbances. By estimating lumped disturbances in fixed time, nominal containment maneuvering control laws are designed in an Earth-fixed reference frame. Input-to-state safe control barrier functions (ISSf-CBFs) are constructed for mapping safety constraints on states to constraints on control inputs. A distributed quadratic optimization problem with the norm of control inputs as the objective function and ISSf-CBFs as constraints is formulated. A recurrent neural network-based neurodynamic optimization approach is adopted to solve the quadratic optimization problem for computing the forces and moments within the safety and input constraints in real time. It is proven that the error signals in the closed-loop control system are uniformly ultimately bounded and the multi-ASVs system is guaranteed for input-to-state safety. Simulation results are elaborated to substantiate the effectiveness of the proposed safety-critical control method for ASVs based on neurodynamic optimization with control barrier functions.

Autonomous marine vehicles (AMVs), including autonomous surface and underwater vehicles, are versatile means to explore, exploit, monitor, and protect marine resources and environments. Motion control is a fundamental enabling technique for state-of-the-art AMV development. Especially, guidance is a critical component in AMV motion control. In recent years, line-of-sight (LOS) guidance, as an efficient guidance method, has attracted tremendous interest from both theoretical and practical perspectives. In this paper, an overview of recent advances in LOS guidance for AMV path following is provided. First, a control objective for the path following of an AMV with a kinematic model is specified. Next, major LOS guidance laws for path following are reviewed in detail. Then, LOS guidance laws applicable to coordinated path following of multiple AMVs are elaborated. Finally, six challenging issues for future research are addressed.

This article investigates the model-free containment control of multiple underactuated unmanned surface vessels (USVs) subject to unknown kinetic models. A novel cooperative control architecture is presented for achieving a containment formation under switching topologies. Specifically, a path-guided distributed containment motion generator (CMG) is first proposed for generating reference points according to the underlying switching topologies. Next, guiding-vector-field-based guidance laws are designed such that each USV can track its reference point, enabling smooth transitions during topology switching. Finally, data-driven neural predictors by utilizing real-time and historical data are developed for estimating total uncertainties and unknown input gains, simultaneously. Based on the learned knowledge from neural predictors, adaptive kinetic control laws are designed and no prior information on kinetic model parameters is required. By using the proposed method, the fleet is able to converge to the convex hull spanned by multiple virtual leaders under switching topologies regardless of fully unknown kinetic models. Through stability analyses, it is proven that the closed-loop control system is input-to-state stable and the tracking errors are uniformly ultimately bounded. Simulation results verify the effectiveness of the proposed cooperative control architecture for multiple underactuated USVs with fully unknown kinetic models.

Cued Speech (CS) is a communication system for deaf people or hearing impaired people, in which a speaker uses it to aid a lipreader in phonetic level by clarifying potentially ambiguous mouth movements with hand shape and positions.Feature extraction of multi-modal CS is a key step in CS recognition.Recent supervised deep learning based methods suffer from noisy CS data annotations especially for hand shape modality.In this work, we first propose a self-supervised contrastive learning method to learn the feature representation of image without using labels.Secondly, a small amount of manually annotated CS data are used to fine-tune the first module.Thirdly, we present a module, which combines Bi-LSTM and self-attention networks to further learn sequential features with temporal and contextual information.Besides, to enlarge the volume and the diversity of the current limited CS datasets, we build a new British English dataset containing 5 native CS speakers.Evaluation results on both French and British English datasets show that our model achieves over 90% accuracy in hand shape recognition.Significant improvements of 8.75% (for French) and 10.09% (for British English) are achieved in CS phoneme recognition correctness compared with the state-of-the-art.

This article addresses the distributed path maneuvering of underactuated unmanned surface vehicles (USVs) with collision avoidance and connectivity preservation. The USVs are guided by the multiple virtual leaders moving along the multiple parameterized paths and only a fraction of USVs have access to the virtual leaders. An observer-based finite-time control method is proposed to achieve a containment formation. Specifically, a finite-time extended state observer is employed to recover the unmeasured linear/angular velocities and estimate the total disturbances consisting of model uncertainties as well as ocean disturbances at first. Then, observer-based finite-time guidance laws using the information of neighbors are designed based on a containment scheme. An artificial potential field is incorporated into the distributed guidance law design to avoid collision and preserve connectivity. Finally, antidisturbance kinetic control laws are devised based on the finite-time convergent observers and nonlinear tracking differentiators. It is proven that all error signals in the closed-loop system are ultimately uniformly bounded, and the distributed path maneuvering formation pattern can be achieved in a finite time when USVs outside the collision avoidance and connectivity preservation region. Simulation results are given to verify the effectiveness of the proposed output feedback control method for the distributed path maneuvering of multiple USVs with position-yaw measurements only.