Search results

The purpose of this study is to reduce the residual stress in welded workpieces, optimize the vibratory stress relief treatment process through the use of a vibration generator and enhance the durability and longevity of the workpiece by developing a vibratory stress relief robot that incorporates a multi-manipulator system.

The multi-manipulator combination work is designed so that each manipulator is deployed according to the requirements of vibration stress relief work. Each manipulator works independently and coordinates with others to achieve multi-dimensional vibratory stress relief of the workpiece. A two-degree-of-freedom mobile platform is designed to enable the transverse and longitudinal movement of the manipulator, expanding the working space of the robot. A small electromagnetic superharmonic vibration generator is designed to produce directional vibrations in any orientation. This design addresses the technical challenge of traditional vibration generators being bulky and unable to achieve directional vibrations.

The residual stress relief experiment demonstrates that the residual stress of the workpiece is reduced by approximately 73% through three-degree-of-freedom vibration. The multi-dimensional vibration effectively enhances the relief effect of residual stress, which is beneficial for improving the strength and service life of the workpiece.

A new multi-manipulator robot is proposed to alleviate the residual stress generated by workpiece welding by integrating vibratory stress relief with robotics. It is beneficial to reduce material and energy consumption while enhancing the strength and service life of the workpiece.

The purpose of this paper is to enhance the performance of robots in peg-in-hole assembly tasks, enabling them to swiftly and robustly accomplish the task. It also focuses on the robot’s ability to generalize across assemblies with different hole sizes.

Human behavior in peg-in-hole assembly serves as inspiration, where individuals visually locate the hole firstly and then continuously adjust the peg pose based on force/torque feedback during the insertion process. This paper proposes a novel framework that integrate visual servo and adjustment based on force/torque feedback, the authors use deep neural network (DNN) and image processing techniques to determine the pose of hole, then an incremental learning approach based on a broad learning system (BLS) is used to simulate human learning ability, the number of adjustments required for insertion process is continuously reduced.

The author conducted experiments on visual servo, adjustment based on force/torque feedback, and the proposed framework. Visual servo inferred the pixel position and orientation of the target hole in only about 0.12 s, and the robot achieved peg insertion with 1–3 adjustments based on force/torque feedback. The success rate for peg-in-hole assembly using the proposed framework was 100%. These results proved the effectiveness of the proposed framework.

This paper proposes a framework for peg-in-hole assembly that combines visual servo and adjustment based on force/torque feedback. The assembly tasks are accomplished using DNN, image processing and BLS. To the best of the authors’ knowledge, no similar methods were found in other people’s work. Therefore, the authors believe that this work is original.

Current reinforcement learning (RL) algorithms are facing issues such as low learning efficiency and poor generalization performance, which significantly limit their practical application in real robots. This paper aims to adopt a hybrid model-based and model-free policy search method with multi-timescale value function tuning, aiming to allow robots to learn complex motion planning skills in multi-goal and multi-constraint environments with a few interactions.

A goal-conditioned model-based and model-free search method with multi-timescale value function tuning is proposed in this paper. First, the authors construct a multi-goal, multi-constrained policy optimization approach that fuses model-based policy optimization with goal-conditioned, model-free learning. Soft constraints on states and controls are applied to ensure fast and stable policy iteration. Second, an uncertainty-aware multi-timescale value function learning method is proposed, which constructs a multi-timescale value function network and adaptively chooses the value function planning timescales according to the value prediction uncertainty. It implicitly reduces the value representation complexity and improves the generalization performance of the policy.

The algorithm enables physical robots to learn generalized skills in real-world environments through a handful of trials. The simulation and experimental results show that the algorithm outperforms other relevant model-based and model-free RL algorithms.

This paper combines goal-conditioned RL and the model predictive path integral method into a unified model-based policy search framework, which improves the learning efficiency and policy optimality of motor skill learning in multi-goal and multi-constrained environments. An uncertainty-aware multi-timescale value function learning and selection method is proposed to overcome long horizon problems, improve optimal policy resolution and therefore enhance the generalization ability of goal-conditioned RL.

The null-space projection method is commonly adopted for controlling redundant robots, which undoubtedly requires the robot Jacobian matrix inverse. This paper aims to provide a novel control scheme, which enables null-space control of redundant robots without conflict with the main task space.

In this paper, an impedance-based null-space control approach for redundant robots is proposed. The null-space degrees of freedom are separated from the primary task space by using the eigenvalue decomposition. Then, a joint impedance controller spans the null space and is reflected into the joint space to manage the redundancy. Finally, several experiments have been conducted to evaluate and validate the performance of the proposed approach in comparison with the null-space projection method under various situations.

Experiment results show that no significant differences were observed between the different filling eigenvalues in the proposed approach under different null-space dimensions and motion velocity. Besides, comparative experiment results demonstrate that the proposed method can achieve comparable results to the null-space projection method. Nevertheless, the suggested approach has benefits regarding the quantity of control parameters in addition to not requiring a Jacobian inverse. Notably, the performance of the proposed method will improve as the null-space dimension increases.

This study presents a new control method for redundant robots, which has advantages for dealing with the problems of controlling redundant robots compared to the existing methods.

Aiming at the problem of insufficient adaptability of robot motion planners under the diversity of end-effector constraints, this paper proposes Transformation Cross-sampling Framework (TC-Framework) that enables the planner to adapt to different end-effector constraints.

This work presents a standard constraint methodology for representing end-effector constraints as a collection of constraint primitives. The constraint primitives are merged sequentially into the planner, and a unified constraint input interface and constraint module are added to the standard sampling-based planner framework. This approach enables the realization of a generic planner framework that avoids the need to build separate planners for different end-effector constraints.

Simulation tests have demonstrated that the planner based on TC-framework can adapt to various end-effector constraints. Physical experiments have also confirmed that the framework can be used in real robotic systems to perform autonomous operational tasks. The framework’s strong compatibility with constraints allows for generalization to other tasks without modifying the scheduler, significantly reducing the difficulty of robot deployment in task-diverse scenarios.

This paper proposes a unified constraint method based on constraint primitives to enhance the sampling-based planner. The planner can now adapt to different end effector constraints by opening up the input interface for constraints. A series of simulation tests were conducted to evaluate the TC-Framework-based planner, which demonstrated its ability to adapt to various end-effector constraints. Tests on a physical experimental system show that the framework allows the robot to perform various operational tasks without requiring modifications to the planner. This enhances the value of robots for applications in fields with diverse tasks.

Collaborative robots (cobots) are widely used in various manipulation tasks within complex industrial environments. However, the manipulation capabilities of cobot manipulation planning are reduced by task, environment and joint physical constraints, especially in terms of force performance. Existing motion planning methods need to be more effective in addressing these issues. To overcome these challenges, the authors propose a novel method named force manipulability-oriented manipulation planning (FMMP) for cobots.

This method integrates force manipulability into a bidirectional sampling algorithm, thus planning a series of paths with high force manipulability while satisfying constraints. In this paper, the authors use the geometric properties of the force manipulability ellipsoid (FME) to determine appropriate manipulation configurations. First, the authors match the principal axes of FME with the task constraints at the robot’s end effector to determine manipulation poses, ensuring enhanced force generation in the desired direction. Next, the authors use the volume of FME as the cost function for the sampling algorithm, increasing force manipulability and avoiding kinematic singularities.

Through experimental comparisons with existing algorithms, the authors validate the effectiveness and superiority of the proposed method. The results demonstrate that the FMMP significantly improves the force performance of cobots under task, environmental and joint physical constraints.

To improve the force performance of manipulation planning, the FMMP introduces the FME into sampling-based path planning and comprehensively considers task, environment and joint physical constraints. The proposed method performs satisfactorily in experiments, including assembly and in situ measurement.

This study aims to propose a force control algorithm based on neural networks, which enables a robot to follow a changing reference force trajectory when in contact with human skin while maintaining a stable tracking force.

Aiming at the challenge of robots having difficulty tracking changing force trajectories in skin contact scenarios, a single neuron algorithm adaptive proportional – integral – derivative online compensation is used based on traditional impedance control. At the same time, to better adapt to changes in the skin contact environment, a gated recurrent unit (GRU) network is used to model and predict skin elasticity coefficients, thus adjusting to the uncertainty of skin environments.

In two robot–skin interaction experiments, compared with the traditional impedance control and robot force control algorithm based on the radial basis function model and iterative algorithm, the maximum absolute force error, the average absolute force error and the standard deviation of the force error are all decreased.

As the training process of the GRU network is currently conducted offline, the focus in the subsequent phase is to refine the network to facilitate real-time computation of the algorithm.

This algorithm can be applied to robot massage, robot B-ultrasound and other robot-assisted treatment scenarios.

As the proposed approach obtains effective force tracking during robot–skin contact and is verified by the experiment, this approach can be used in robot–skin contact scenarios to enhance the accuracy of force application by a robot.

The study aims to equip robots with the ability to precisely maintain interaction forces, which is crucial for tasks such as polishing in highly dynamic environments with unknown and varying stiffness and geometry, including those found in airplane wings or thin, soft materials. The purpose of this study is to develop a novel adaptive force-tracking admittance control scheme aimed at achieving a faster response rate with higher tracking accuracy for robot force control.

In the proposed method, the traditional admittance model is improved by introducing a pre-proportional-derivative controller to accelerate parameter convergence. Subsequently, the authors design an adaptive law based on fuzzy logic systems (FLS) to compensate for uncertainties in the unknown environment. Stability conditions are established for the proposed method through Lyapunov analysis, which ensures the force tracking accuracy and the stability of the coupled system consisting of the robot and the interaction environment. Furthermore, the effectiveness and robustness of the proposed control algorithm are demonstrated by simulation and experiment.

A variety of unstructured simulations and experimental scenarios are designed to validate the effectiveness of the proposed algorithm in force control. The outcomes demonstrate that this control strategy excels in providing fast response, precise tracking accuracy and robust performance.

In real-world applications spanning industrial, service and medical fields where accurate force control by robots is essential, the proposed method stands out as both practical and straightforward, delivering consistently satisfactory performance across various scenarios.

This research introduces a novel adaptive force-tracking admittance controller based on FLS and validated through both simulations and experiments. The proposed controller demonstrates exceptional performance in force control within environments characterized by unknown and varying.