Industrial robots are widely utilized in the machining of complex parts because of their flexibility. However, their low positioning accuracy and spatial geometric error characteristics significantly limit the contour precision of robot machined parts. Therefore, in the robot machining procedure, an in situ measurement system is typically required. This study aims to enhance the trajectory accuracy of robotic machining through robotic in situ measurement and meta-heuristic optimization. In this study, a measurement-machining dual-robot system for measurement and machining is established, consisting of a measurement robot with a laser sensor mounted at the robot end and a machining robot equipped with a machining tool. In the measuring process, high-precision standard spheres are set on the edge of the machining area, and the high-precision standard geometry is measured by the measurement robot. According to measured geometry information in the local area, the trajectory accuracy for the machining robot is improved. By utilizing the standard radius of the standard spheres and adopting a meta-heuristic optimization algorithm, this study addresses the complexity of the robot kinematics model, while also overcoming local optima commonly introduced by gradient-based iterative methods. The results of the experiments in this study confirm that the proposed method markedly refines the precision of the robot machining trajectory.

Capturing the non-cooperative tumbling target by the free-floating space robot stands as a crucial task within on-orbit servicing. However, the strong dynamic coupling of the base-spacecraft and the manipulator seriously disturbs the base-spacecraft, which reduces the power generation efficiency of solar panels and the communication quality with the earth station. In this paper, the trajectory planning method of the free-floating space robot for non-cooperative tumbling target capture based on deep reinforcement learning is proposed, which can reduce the disturbance of the base-spacecraft during target capture. First, the generalized Jacobian matrix of the space robot is derived, from which the dynamics model is obtained. The kinematics model of the space non-cooperative tumbling target is established. And the contact collision dynamics between the space robot and the tumbling target are analysed. Second, the twin delayed deep deterministic policy gradient algorithm is introduced to plan the trajectory for capturing the non-cooperative tumbling target, where apart from the motion parameters of the manipulator and the generalized manipulability of the space robot, the pose disturbance of the base-spacecraft is initially added to the reward function. Finally, the simulation for target capture is carried out. The results show that compared with the existing method, the proposed method converges faster with a larger reward, and the pose disturbance of the base-spacecraft is reduced. Moreover, the method performs well for capturing the non-cooperative tumbling target with different initial rotational angular velocities.

In this paper, a novel fast nonsingular integral terminal sliding mode controller based on an adaptive neural network (ANN-FNITSMC) is proposed for the trajectory tracking control of cable-driven continuum robots (CDCRs) in complex underwater environments with uncertainties. First, a novel fast nonsingular integral terminal sliding mode control (FNITSMC) is designed to solve the chattering and singularity problems of the conventional terminal sliding mode control (TSMC). Second, an adaptive neural network (ANN) based on a radial basis function (RBF) is established to derive the uncertainties and compensate for the control input of CDCRs, enabling high-stable accuracy and strong robustness trajectory tracking in complex underwater environments. Simulation results are presented to demonstrate the high accuracy and strong robustness of the ANN-FNITSMC. Finally, the high accuracy, high stability, and strong robustness of the proposed trajectory tracking strategy are verified through an underwater experiment platform.

The underwater target detection is affected by image blurring caused by suspended particles in water bodies and light scattering effects. To tackle this issue, this paper proposes a reparameterized feature enhancement and fusion network for underwater blur object recognition (REFNet). First, this paper proposes the reparameterized feature enhancement and gathering (REG) module, which is designed to enhance the performance of the backbone network. This module integrates the concepts of reparameterization and global response normalization to enhance the network’s feature extraction capabilities, addressing the challenge of feature extraction posed by image blurriness. Next, this paper proposes the cross-channel information fusion (CIF) module to enhance the neck network. This module combines detailed information from shallow features with semantic information from deeper layers, mitigating the loss of image detail caused by blurring. Additionally, this paper replace the CIoU loss function with the Shape-IoU loss function improves target localization accuracy, addressing the difficulty in accurately locating bounding boxes in blurry images. Experimental results indicate that REFNet achieves superior performance compared to state-of-the-art methods, as evidenced by higher mAP scores on the underwater robot professional competitionand detection underwater objects datasets. REFNet surpasses YOLOv8 by approximately 1.5% in $mAP_{50:95}$ on the URPC dataset and by about 1.3% on the DUO dataset. This enhancement is achieved without significantly increasing the model’s parameters or computational load. This approach enhances the precision of target detection in challenging underwater environments.

In this paper, fractional-order (FO), intelligent, and robust sliding mode control (SMC) and stabilization of inherently nonlinear, multi-input, multi-output 6-DOF robot manipulators are investigated. To ensure robust control and better performance of the robot system, significant studies on various control transactions have been explored. First, a sliding proportional-integral-derivative (PID) surface is conceived and then its FO constitute is developed. It is an important fact that in SMC, the reaching phase is fast and the chattering is abated in the sliding phase. In particular, the discontinuity in the SMC is prevented in view of the boundary layer obtained by recommending the sigmoid function together with fuzzy logic to eliminate the chattering phenomenon. A hybrid tuning method consisting of gray wolf optimization and particle swarm optimization (GWO-PSO) algorithms is applied to tune the parameters of PID sliding mode control (PIDSMC), FO PIDSMC (FOPIDSMC), fuzzy PIDSMC (FPIDSMC), and FO fuzzy PIDSMC (FOFPIDSMC) controllers. In simulation results, the tuned FOFPIDSMC controller consistently outperforms PIDSMC, FOPIDSMC, and FPIDSMC controllers tuned by the GWO-PSO in dynamic performance, trajectory tracking, disturbance rejection, and mass uncertainty scenarios. It has been seen through a thorough performance analysis that 91.93% and 44.13% improvement are, respectively, obtained for mean absolute error (MAE) and torques root mean square (RMS) values of the joints when using from the PIDSMC to the FOFPIDSMC. Finally, the simulation outcomes reveal the superior aspects of the designed FOFPIDSMC and also demonstrate that the FOFPIDSMC controller enhances the dynamic performances of the 6-revolute universal robots 5 (6R UR5) robot manipulator under a variety of operating conditions.

Adaptation to the dynamic environment and variable task sequence is the critical ability for robot navigation and task execution. The Cyclic Networking Rapidly-exploring Random Tree (CNRRT) method is proposed to obtain the optimal path in real time and realize long-term path planning ability in a complex dynamic environment. The cyclic branch is introduced to the acyclic graph of Rapidly-exploring Random Tree (RRT) method, which forms a decentralized path network in the configuration space. An iterative searching strategy is built to search for the optimal path in the network. The branch prune, reconnection, and regrowth processes enable the decentralized network to efficiently respond to dynamic changes in the environment. The CNRRT can search for the real-time optimal path in the dynamic environment, dealing with the configuration and task changes robustly. Besides, the CNRRT is consistent for scenarios with long-term task sequence without significant performance fluctuation. Simulations and real-world comparative experiments verify the effectiveness of the proposed method.

In recent years, there has been growing interest in developing robots and autonomous systems that can interact with human in a more natural and intuitive way. One of the key challenges in achieving this goal is to enable these systems to manipulate objects and tools in a manner that is similar to that of humans. In this paper, we propose a novel approach for learning human-style manipulation skills by using adversarial motion priors, which we name HMAMP. The approach leverages adversarial networks to model the complex dynamics of tool and object manipulation and the aim of the manipulation task. The discriminator is trained using a combination of real-world data and simulation data executed by the agent, which is designed to train a policy that generates realistic motion trajectories that match the statistical properties of human motion. We evaluated HMAMP on one challenging manipulation task: hammering, and the results indicate that HMAMP is capable of learning human-style manipulation skills that outperform current baseline methods. Additionally, we demonstrate that HMAMP has potential for real-world applications by performing real robot arm hammering tasks. In general, HMAMP represents a significant step towards developing robots and autonomous systems that can interact with humans in a more natural and intuitive way, by learning to manipulate tools and objects in a manner similar to how humans do.

In scenarios such as environmental data collection and traffic monitoring, timely responses to real-time situations are facilitated by persistently accessing nodes with revisiting constraints using unmanned aerial vehicles (UAVs). However, imbalanced task allocation may pose risks to the safety of UAVs and potentially lead to failures in monitoring tasks. For instance, continuous visits to nodes without replenishment may damage UAV batteries, while delays in recharging could result in missing task deadlines, ultimately causing task failures. Therefore, this study investigates the problem of achieving balanced multi-UAV path planning for persistent monitoring tasks, which has not been previously researched according to the authors’ knowledge. The main contribution of this study is the proposal of two novel indicators to assist in balancing task allocation regarding multi-UAV path planning for persistent monitoring. One of the indicators is namely the waiting factor, which reflects the urgency of a task node waiting to be accessed, and the other is the difficulty level which is introduced to measure the difficulty of tasks undertaken by a UAV. By minimizing differences in difficulty level among UAVs, we can ensure equilibrium in task allocation. For a single UAV, the ant colony initialized genetic algorithm (ACIGA) has been proposed to plan its path and obtain its difficulty level. For multiple UAVs, the K-means clustering algorithm has been improved based on difficulty levels to achieve balanced task allocation. Simulation experiments demonstrated that the difficulty level could effectively reflect the difficulty of tasks and that the proposed algorithms could enable UAVs to achieve balanced task allocation.

In recent years, the removal of orbital debris has become an increasingly urgent task due to advancements in human space exploration. Capturing space debris through caging manipulation offers notable advantages in terms of robustness and controllability. This paper proposes a configuration-based caging manipulation design method for a cable-driven flexible arm. First, the cable-driven flexible arm with multi self-lockable links is introduced. To quantify the caging configurations formed by different self-lockable link selections, a novel planar caging quality metric is then presented for arbitrary planar objects. Using this metric, a caging design method is developed for the flexible arm to conduct caging manipulations. Finally, the caging manipulation strategies are discussed with lock selection maps for different objects, and a robust caging strategy considering uncertainty is further explored. Simulation and experimental results validate the effectiveness of the proposed caging design method and demonstrate better performance compared to conventional caging methods.

Biped wall-climbing robots (BWCRs) serve as viable alternatives to human workers for inspection and maintenance tasks within three-dimensional (3D) curtain wall environments. However, autonomous climbing in such environments presents significant challenges, particularly related to localization and navigation. This paper presents a pioneering navigation framework tailored for BWCRs to navigate through 3D curtain wall environments. The framework comprises three essential stages: Building Information Model (BIM)-based map extraction, 3D climbing path planning (based on our previous work), and path tracking. An algorithm is developed to extract a detailed 3D map from the BIM, including structural elements such as walls, frames, and ArUco markers. This generated map is input into a proposed path planner to compute a viable climbing motion. For path tracking during actual climbing, an ArUco marker-based global localization method is introduced to estimate the pose of the robot, enabling adjustments to the target foothold by comparing desired and actual poses. The conducted experiments validate the feasibility and efficacy of the proposed navigation framework and associated algorithms, aiming to enhance the autonomous climbing capability of BWCRs.

The selection of random sampling points is crucial for the path quality generated by probabilistic roadmap (PRM) algorithm. Increasing the number of sampling points can enhance path quality. However, it may also lead to extended convergence time and reduced computational efficiency. Therefore, an improved probabilistic roadmap algorithm (TL-PRM) is proposed based on topological discrimination and lazy collision. TL-PRM algorithm first generates a circular grid area among start and goal points. Then, it constructs topological nodes. Subsequently, elliptical sampling areas are created between each pair of adjacent topological nodes. Random sampling points are generated within these areas. These sampling points are interconnected using a layer connection strategy. An initial path is generated using a delayed collision strategy. The path is then adjusted by modifying the nodes on the convex outer edges to avoid obstacles. Finally, a reconnection strategy is employed to optimize the path. This reduces the number of path waypoints. In dynamic environments, TL-PRM algorithm employs pose adjustment strategies for semi-static and dynamic obstacles. It can use either the same or opposite pose adjustments to avoid dynamic obstacles. Experimental results indicate that TL-PRM algorithm reduces the average number of generated sampling points by 70.9% and average computation time by 62.1% compared with PRM* and PRM-Astar algorithms. In winding and narrow passage maps, TL-PRM algorithm significantly decreases the number of sampling points and shortens convergence time. In dynamic environments, the algorithm can adjust its pose orientation in real time. This allows it to safely reach the goal point. TL-PRM algorithm provides an effective solution for reducing the generation of sampling points in PRM algorithm.

This study introduces an innovative deep learning method for intelligent healthcare emotion analysis, specifically targeting the recognition of pain based on facial expressions. The suggested approach combines cloud-based mobile applications, utilising separate front-end and back-end elements to optimise data processing. The main contributions consist of a Smart Automated System (SASys) that integrates statistical and deep learning methods to extract features, thereby guaranteeing both resilience and efficiency. Image preprocessing encompasses the tasks of detecting faces and normalising them, which is crucial for extracting features with high accuracy. The comparison of statistical feature representation using Histogram of Oriented Gradients and Local Binary Pattern, along with machine learning classifiers, against an enhanced deep learning-based approach with an integrated multi-tasking feature known as multi-task convolutional neural network, demonstrates encouraging outcomes that support the superiority of the convolutional neural network architecture. Statistical and deep learning-based classification scores, when combined, greatly enhance the system’s overall performance. The results of the experiments prove that the method is effective, outperforming traditional classifiers and exhibiting comparable accuracy to cutting-edge healthcare SASys.

Spherical robots face significant challenges in motion control on non-horizontal terrains, such as slopes, due to their unique spherical structure. This paper systematically investigates the motion stability of spherical robots on inclined surfaces through modeling, control algorithm design, and experimental validation. Precise Equilibrium Modeling: Using the virtual displacement method, the precise equilibrium equation for spherical robots on slopes is derived, addressing the issue of insufficient accuracy in describing the actual center of gravity in existing studies. Control Algorithm Design: For known slope conditions, a Backstepping Control (BSC) algorithm is designed, demonstrating excellent tracking performance. For unknown slope conditions, an Adaptive Backstepping Control (ABSC) algorithm is proposed, which significantly reduces tracking errors and enhances system robustness through parameter adaptation. Simulation and Physical Validation: Simulations confirm the effectiveness of the algorithms: BSC achieves high-precision control under known slopes, while ABSC exhibits strong adaptability under unknown slopes. Physical experiments validate the stability of the algorithms in a $5^\circ$ slope environment, demonstrating reliable performance across different control angles.

Indoor positioning systems (IPS) are essential for mobile robot navigation in environments where global positioning systems (GPS) are unavailable, such as hospitals, warehouses, and intelligent infrastructure. While current surveys may limit themselves to specific technologies or fail to provide practical application-specific details, this review summarizes IPS developments directed specifically towards mobile robotics. It examines and compares a breadth of approaches that vary across non-radio frequency, radio frequency, and hybrid sensor fusion systems, through the lens of performance metrics that include accuracy, delay, scalability, and cost. Distinctively, this work explores emerging innovations, including synthetic aperture radar (SAR), federated learning, and privacy-aware AI, which are reshaping the IPS landscape. The motivation stems from the’ increasing complexity and dynamic nature of indoor environments, where high-precision, real-time localization is essential for safety and efficiency. This literature review provides a new conceptual, cross-border pathway for research and implementation of IPS in mobile robotics, addressing both technical and application-related challenges in sectors related to healthcare, industry, and smart cities. The findings from the literature review allow early career researchers, industry knowledge workers, and stakeholders to provide secure societal, human, and economic integration of IPS with AI and IoT in safe expansions and scale-ups.

Extensible continuum robots (ECRs) offer distinct advantages over conventional continuum robots due to their ability to enhance workspace adaptability through length adjustments. This makes ECRs particularly promising for applications that require variable lengths involving the manipulation of objects in challenging environments, such as risky, cluttered, or confined. The development of ECRs necessitates careful consideration of mechanical structures, actuation methods, methods of stiffness variability, and control methods. The selection of papers is based on their relevance to ECRs within the period of 2010 to 2023 in the databases of Scopus and Web of Science. Distinguishing itself from other review papers, this paper aims to deliver a comprehensive and critical discussion about the advantages and disadvantages of ECRs concerning their mechanical structures, actuation methods, stiffness variability, and control methods. It is a beneficial resource for researchers and engineers interested in ECRs, providing essential insights to guide future developments in this field. Based on the literature, existing ECRs exhibit an inherent trade-off between flexibility and structural strength due to the absence of systematic design methods. Additionally, there is a lack of intelligent and effective controllers for achieving complex control performance and autonomous stiffness variability.

With advancements in industrial robot technology and the ongoing enhancements in control system performance, the demand for precise robot motion is increasing. Generally, an increased number of interpolation points enhances the precision of robot movement, but excessive points can lead to jittering and out-of-step issues. This paper investigates the relationship between the number of motion interpolation points and the response times of the control system and the robot’s terminal velocity, based on the theoretical calculation and experimental analysis of the limit interpolation points for the control system of a self-developed 6-DOF (Six Degree of Freedom) robot. The method for calculating limit interpolation points is refined using the least squares method, and equations are derived for different control system response time and robot’s terminal velocity reaction times. The validity of the prediction curves is verified through experimental analysis.

The inverse dynamics model of an industrial robot can predict and control the robot’s motion and torque output, improving its motion accuracy, efficiency, and adaptability. However, the existing inverse rigid body dynamics models still have some unmodelled residuals, and their calculation results differ significantly from the actual industrial robot conditions. The bootstrap aggregating (bagging) algorithm is combined with a long short-term memory network, the linear layer is introduced as the network optimization layer, and a compensation method of hybrid inverse dynamics model for robots based on the BLL residual prediction algorithm is proposed to meet the above needs. The BLL residual prediction algorithm framework is presented. Based on the rigid body inverse dynamics of the Newton–Euler method, the BLL residual prediction network is used to perform error compensation on the inverse dynamics model of the Franka robot. The experimental results show that the hybrid inverse dynamics model based on the BLL residual prediction algorithm can reduce the average residuals of the robot joint torque from 0.5651 N·m to 0.1096 N·m, which improves the accuracy of the inverse dynamics model compared with those of the rigid body inverse dynamics model. This study lays the foundation for performing more accurate operation tasks using industrial robots.

The problem of how to effectively track and intercept small aircraft that break into the no-fly zones is now attracting increasing interest in robotics society. Vision-based control has been proved an effective solution to the target tracking problem for unmanned aerial vehicles (UAVs). Due to the limited field of view (FOV) of onboard vision sensors, existing works assume that the target is always detectable during tracking or limit the flight speed of the UAV in practice. In this paper, inspired by the broad FOV of camera network, we are the first to propose an eye-to-hand (i.e., fixed cameras) visual servoing scheme to track and intercept aerial targets by using UAVs and ground visual sensors. Specifically, utilizing rotation matrices, we first present a visual servoing equation to convert the UAV motion in image planes to the inertial frame. Then, an image-based visual servoing controller is designed directly based on image errors of camera nodes in the sensor network, and system stability is proved by means of Lyapunov analysis. Additionally, to achieve the desired translational velocity command, a low-level attitude controller is developed based on the UAV dynamics. Finally, a series of experiments in both simulated and real flight scenarios show the outstanding efficacy of our method.

A more intuitive appreciation of spatial compliant behavior can be obtained through analysis and description of the behavior in terms of its centers, specifically the center of stiffness, the center of compliance, and the center of elasticity. This paper investigates the properties of each of these centers. Necessary and sufficient conditions for the coincidence of these centers are identified. A physical appreciation of those compliant behaviors that have coincident centers is obtained in terms of restrictions on the geometry of topologically simple mechanisms that realize those behaviors. The results can be used in the design of compliant mechanisms for robotic manipulation, especially when the compliance is characterized by the location of its center.