Explore the vast range of titles available.

This paper considers the problem of the accurate task space finite-time control susceptible to both undesirable disturbance forces exerted on the end-effector and unknown friction forces coming from joints directly driven by the actuators as well as unstructured forces resulting from the kinematic singularities appearing on the mechanism trajectory. We obtain a class of estimated extended transposed Jacobian controllers which seem to successfully counteract the external disturbance forces on the basis of a suitably defined task-space non-singular terminal sliding manifold (TSM) and the Lyapunov stability theory. Moreover, in order to overcome (or to minimise) the undesirable chattering effects, the proposed robust control law involves the second-order sliding technique. The numerical simulations (closely related to an experiment) ran for a mobile manipulator consisting of a non-holononic platform of (2;0) type and a holonomic manipulator of two revolute kinematic pairs show the performance of the proposed controllers and make a comparison with other well-known control schemes.

This paper presents a robust nonlinear controller for a 6 degree of freedom (DOF) parallel manipulator in the task space coordinates. The proposed control strategy requires information on orientations and translations in the task space unlike the joint space or link space control scheme. Although a 6 DOF sensor may provide such information in a straightforward manner, its cost calls for a more economical alternative. A novel indirect method based on the readily available length information engages as a potential candidate to replace a 6 DOF sensor. The indirect approach generates the necessary information by solving the forward kinematics and subsequently applying alpha-beta-gamma tracker. With the 6 DOF signals available, a robust nonlinear task space control (RNTC) scheme is proposed based on the Lyapunov redesign method, whose stability is rigorously proved. The performance of the proposed RNTC with the new estimation scheme is evaluated via experiments. First, the results of the estimator are compared with the rate-gyro signals, which indicates excellent agreement. Then, the RNTC with on-line estimated 6 DOF data is shown to achieve excellent control performance to sinusoidal inputs, which is superior to those of a commonly used proportional-plus-integral-plus-derivative controller with a feedforward friction compensation under joint space coordinates and the nonlinear controller under task space coordinates.

During the last two decades, parallel robots have become more ubiquitous, employed in a great variety of sectors, from food to aerospace industries. In fact, they are much more efficient than their serial counterparts in terms of performing fast motions and consuming less energy. However, due to their mechanical complexity, they present a highly complex non-linear dynamics, which makes the modelling and control tasks difficult. Aiming to improve the performance and robustness of the control laws already used to control this type of mechanisms, previously, the authors proposed two novel laws of hybrid control, implemented in the joint space, in order to improve the dynamic behavior of parallel robots when performing fast motion tasks. Among the goals of the current work, one can mention to adapt the two laws of hybrid control, proposed in the previous work, by implementing them in the task space. Additionally, the peculiarities related to the dynamic formulation and the tuning of controller gains are also shown. Furthermore, a comparison of the performances of the pure and hybrid control techniques, implemented in both joint and task spaces, is presented as well, by executing the same paths and using adequate metrics. In the selected paths, experiments revealed that the hybridization process of pure control laws in the task space provides a significant reduction of the path-tracking and steady-state errors.

In this study, task space tracking control of robot manipulators driven by brushless DC (BLDC) motors is considered. Dynamics of actuators are taken into account and the entire electromechanical system (i.e. kinematic, dynamic, and electrical models) is assumed to include parametric/structured uncertainties. A novel adaptive controller is designed and the stability of the closed loop system is ensured via novel Lyapunov type tools. To demonstrate performance and applicability of the proposed method, a simulation study is conducted using the model of a two degree of freedom, planar robotic manipulator driven by BLDC motors. This paper presents a backstepping‐based adaptive controller that is able to handle the parametric uncertainties in the entire electromechanical model of robot manipulators driven by BLDC motors and does not require measurements of joint accelerations. Despite the uncertainties, boundedness of all signals under closed loop operation and global asymptotic stability of task space tracking error are guaranteed by the application of novel Lyapunov type synthesis and stability analysis methods.

The growing interest in the flexibility and operational capabilities of soft manipulators in confined spaces emphasizes the need for precise modeling and accurate motion control. Conventional control methods encounter difficulties in modeling and involve intricate computations. This work introduces a novel deep reinforcement learning (DRL) control algorithm based on neural network modeling. Using the Whale Optimization Algorithm, an approximate dynamic model for the soft manipulator is established. The twin delayed deterministic policy gradient is employed for DRL control. Domain randomization is applied during pretraining in a simulated environment. The algorithm addresses issues related to dependency on measurement data quality and redundant mappings, outperforming other methods by 8–15 mm in control accuracy. The trained DRL controller achieves precise trajectory tracking within the soft manipulator's task space, enabling successful grasping tasks in various complex environments, including pipelines and other narrow spaces. Experimental results confirm the autonomy of our controller in performing these tasks without human intervention. This work proposes a modified twin delayed deterministic policy gradient algorithm in combination with long short‐term memory neural networks to control a soft manipulator. Multiscenario experiments are carried out to validate its effectiveness, such as pipeline operation and delicate object grasping.

The robust dynamic walking control of humanoid robots is the foundation for their application in complex scenarios. Model Predictive Control (MPC) can predict the robot’s motion state over a future time horizon to obtain optimal control inputs and has achieved significant success in walking control of legged robots. However, the high computational cost makes it challenging to ensure real-time performance when dealing with complex dynamic systems. A novel dynamic walking control framework for humanoid robots is introduced in this paper by integrating an efficient footstep planner and Task Space Inverse Dynamics based Whole Body Controller (TSID-WBC). This framework aims to enhance the dynamic performance and robustness of the humanoid robot’s walking control. Firstly, design a bionic step time generator based on relevant research in biology. Then, the MPC controller, combined with the differential equation solution of the Linear Inverted Pendulum (LIP) dynamics model, is used to predict the optimal footsteps for the humanoid robot. Finally, TSID-WBC is employed to optimize joint torques, joint accelerations, and generalized contact forces based on predefined task priorities. The proposed framework avoids using MPC for predictive control of complex dynamic models. Instead, it innovatively integrates the bionic step time generator proposed in this paper with an MPC based on the differential solution of the LIP dynamics model, resulting in an efficient footstep planner. The simulation and experimental results of the humanoid robot Kuavo demonstrate that the proposed control framework achieves better tracking performance for desired walking speeds and responds quickly to external disturbances.

The bilateral teleoperation system is susceptible to model parameter uncertainty and unknown disturbances in both the master and slave manipulators, resulting in instability and inaccuracies in the force and position tracking performance. To address these issues, a novel nonlinear model reference adaptive impedance controller has been designed to achieve coordinated force and position synchronization of dual manipulators. The adaptive control laws, based on sliding mode functions, have been designed to compensate for the uncertainty of the manipulator model. Furthermore, an adaptive estimation law has been employed to appraise the unknown upper bound of external disturbances. This ensures that the closed-loop model parameters of the dual manipulator converge to the reference impedance model respectively. Simultaneously, it enables the position error between the reference model response and the end-effector task space position to asymptotically converge to zero. To verify the effectiveness of the proposed controller, simulations have been conducted on the MATLAB platform and experiments on a single degree of freedom teleoperation system have been performed. The results demonstrate that the controller exhibits strong robustness and has the capability of force-position tracking ability.

In this paper, we propose a set of robust training methods for deep reinforcement learning to transfer learning acquired in one control task to a set of previously unseen control tasks. We improve generalization in commonly used transfer learning benchmarks by a novel sample elimination technique, early stopping, and maximum entropy adversarial reinforcement learning. To generate robust policies, we use sample elimination during training via a method we call strict clipping. We apply early stopping, a method previously used in supervised learning, to deep reinforcement learning. Subsequently, we introduce maximum entropy adversarial reinforcement learning to increase the domain randomization during training for a better target task performance. Finally, we evaluate the robustness of these methods compared to previous work on simulated robots in target environments where the gravity, the morphology of the robot, and the tangential friction coefficient of the environment are altered.

Uncertainty and Jacobian transformation matrix (JMT) are two critical aspects that affect the coordinated robotic arm systems (CRAS) to achieve high accuracy task space trajectory tracking. To address the above two problems, a robust control design and parameter optimization method is proposed for the task space trajectory tracking of the CRAS in this paper. First, a fuzzy dynamical model of the CRAS is established. In this model, the uncertainty is assumed to be bounded and described by fuzzy set theory. It provides a bridge between the dynamical model and the practical system. Then, based on the fuzzy dynamical model, a robust approximate constraint-following servo control is developed to guarantee uniform boundedness (UB) and uniform ultimate boundedness (UUB) of the controlled CRAS. The proposed control can realize the trajectory tracking in task space without JMT, which alleviates the difficulty of control design and implementation. Third, the optimal parameter of the proposed control is selected by solving a fuzzy-based performance index. This performance index is formulated to merge the system manifestation and the control consumption. Finally, a numerical simulation of a dual-arm system is carried out to show the effectiveness of the proposed control method.

PurposeThe purpose of this paper is to solve the tracking problem for free-floating space manipulators (FFSMs) in task space with parameter uncertainties and external disturbance.Design/methodology/approachIn this paper, the novel cerebellar model articulation controller (CMAC) is designed with the feedback controller. More precisely, the parameter uncertainties in the FFSM are considered for achieving the robustness.FindingsBy using the dynamically equivalent model, the CMAC can be designed and trained with the desired performance, such that the prescribed trajectory can be followed accordingly. The simulation results are presented for illustrating the validity of the derived results.Originality/valueBased on the designed CMAC, the tracking error would be approaching zero by choosing appropriate quantization level in CMAC and the corresponding learning rules can be tuned online.