Search results

This paper aims to introduce an imitation learning framework for a wheeled mobile manipulator based on dynamical movement primitives (DMPs). A novel mobile manipulator with the capability to learn from demonstration is introduced. Then, this study explains the whole process for a wheeled mobile manipulator to learn a demonstrated task and generalize to new situations. Two visual tracking controllers are designed for recording human demonstrations and monitoring robot operations. The study clarifies how human demonstrations can be learned and generalized to new situations by a wheel mobile manipulator.

The kinematic model of a mobile manipulator is analyzed. An RGB-D camera is applied to record the demonstration trajectories and observe robot operations. To avoid human demonstration behaviors going out of sight of the camera, a visual tracking controller is designed based on the kinematic model of the mobile manipulator. The demonstration trajectories are then represented by DMPs and learned by the mobile manipulator with corresponding models. Another tracking controller is designed based on the kinematic model of the mobile manipulator to monitor and modify the robot operations.

To verify the effectiveness of the imitation learning framework, several daily tasks are demonstrated and learned by the mobile manipulator. The results indicate that the presented approach shows good performance for a wheeled mobile manipulator to learn tasks through human demonstrations. The only thing a robot-user needs to do is to provide demonstrations, which highly facilitates the application of mobile manipulators.

The research fulfills the need for a wheeled mobile manipulator to learn tasks via demonstrations instead of manual planning. Similar approaches can be applied to mobile manipulators with different architecture.

Proper platform pose is important for the mobile manipulator to accomplish dexterous manipulation tasks efficiently and safely, and the evaluation criterion to qualify manipulation performance is critical to support the pose decision process. This paper aims to present a comprehensive index to evaluate the manipulator’s operation performance from various aspects.

In this research, a criterion called hybrid manipulability (HM) is proposed to assess the performance of the manipulator’s operation, considering crucial factors such as joint limits, obstacle avoidance and stability. The determination of the optimal platform pose is achieved by selecting the pose that maximizes the HM within the feasible inverse reachability map associated with the target object.

A self-built mobile manipulator is adopted as the experimental platform, and the feasibility of the proposed method is experimentally verified in the context of object-grasping tasks both in simulation and practice.

The proposed HM extends upon the conventional notion of manipulability by incorporating additional factors, including the manipulator’s joint limits, the obstacle avoidance situation during the operation and the manipulation stability when grasping the target object. The manipulator can achieve enhanced stability during grasping when positioned in the pose determined by the HM.

Mobile manipulators offer great capability, but their teleoperation is often an overwhelming task for humans due to the many degrees‐of‐freedom of control available from both the mobile platform and the associated manipulator. The purpose of this paper is to address the question of how these controls should be mapped to the robotic mobile platform and its manipulator for “optimal teleoperation”, for the special case of an omnidirectional mobile platform and two joint (with wrist) planar manipulator.

In this paper, the authors summarize the results of a study to optimize the teleoperation interface for a two‐link planar manipulator with a wrist that was mounted on an omni‐directional mobile platform.

The research comprised a carefully‐controlled study using 33 human subjects in seven different treatments of possible control interfaces.

Users performed movement and manipulation tasks, and their performance was measured on several scales.

Based on this study, the authors present guidelines for optimizing mobile manipulator control interfaces and motivate future research using the method of controlled multi‐user trials.

This research has the potential to guide the improvement of interfaces for mobile robots in military, service, and security applications.

The value of this research extends to optimizing remote control schemes to relieve operator fatigue and optimize interface design.

The purpose of this paper is to present the radial basis function (RBF) networks‐based adaptive robust control for an omni‐directional wheeled mobile manipulator in the presence of uncertainties and disturbances.

First, a dynamic model is obtained based on the practical omni‐directional wheeled mobile manipulator system. Second, the RBF neural network is used to identify the unstructured system dynamics directly due to its ability to approximate a nonlinear continuous function to arbitrary accuracy. Using the learning ability of neural networks, RBFNARC can co‐ordinately control the omni‐directional mobile platform and the mounted manipulator with different dynamics efficiently. The implementation of the control algorithm is dependent on the sliding mode control.

Based on the Lyapunov stability theory, the stability of the whole control system, the boundedness of the neural networks weight estimation errors, and the uniformly ultimate boundedness of the tracking error are all strictly guaranteed.

In this paper, an adaptive robust control scheme using neural networks combined with sliding mode control is proposed for crawler‐type mobile manipulators in the presence of uncertainties and disturbances. RBF neural networks approximate the system dynamics directly and overcome the structured uncertainty by learning. Based on the Lyapunov stability theory, the stability of the whole control system, the boundedness of the neural networks weight estimation errors, and the uniformly ultimate boundedness of the tracking error are all strictly guaranteed.

When the mobile manipulator is traveling on an unconstructed terrain, the external disturbance is generated. The load on the end of the mobile manipulator will be affected strictly by the disturbance. The purpose of this paper is to reject the disturbance and keep the end effector in a stable pose all the time, a control method is proposed for the onboard manipulator.

In this paper, the kinematics and dynamics models of the end pose stability control system for the tracked robot are built. Through the guidance of this model information, the control framework based on active disturbance rejection control (ADRC) is designed, which keeps the attitude of the end of the manipulator stable in the pitch, roll and yaw direction. Meanwhile, the control algorithm is operated with cloud computing because the research object, the rescue robot, aims to be lightweight and execute work with remote manipulation.

The challenging simulation experiments demonstrate that the methodology can achieve valid stability control performance in the challenging terrain road in terms of robustness and real-time.

This research facilitates the stable posture control of the end-effector of the mobile manipulator and maintains it in a suitable stable operating environment. The entire system can normally work even in dynamic disturbance scenarios and uncertain nonlinear modeling. Furthermore, an example is given to guide the parameter tuning of ADRC by using model information and estimate the unknown internal modeling uncertainty, which is difficult to be modeled or identified.

The purpose of this paper is to address the trajectory tracking control in task space of a non‐holonomic wheeled mobile manipulator with parameter uncertainties and disturbances. The proposed algorithm is robust adaptive control strategy where parametric uncertainties are compensated by adaptive update techniques and the disturbances are suppressed. The system stability and the convergence of tracking errors to zero are rigorously proved using a Lyapunov theory.

The proposed algorithm is derived based on the advantage of the robot regressor dynamics that express the highly non‐linear robot dynamics in a linear form in terms of the known and unknown robot parameters. The update law for the unknown dynamic parameters is obtained using Lyapunov theory.

Simulation experiments show the effectiveness of the proposed robust adaptive based controller in comparison with a classical passivity based controller.

The proposed adaptive approach is interesting for the control of the mobile manipulators in the task space coordinate even in the presence of dynamic uncertainties and external disturbances.

The purpose of this paper is to determine the base position and the largest working area for mobile manipulators. The base position determines the workspace of the mobile manipulator, particularly when the operation mode is intermittent (i.e. the mobile platform stops when the manipulator conducts the task). When the base of the manipulator is in the intersection area of the Base’s Workable Location Spaces (BWLSes), the end effector (EE) can reach all path points. In this study, the intersection line of BWLSes is calculated numerically, and the largest working area is determined using the BWLS concept. The performance of this method is validated with simulations on specific surface segments, such as plane, cylinder and conical surface segments.

The BWLS is used to determine the largest working area and the base position in which the mobile manipulator can reach all path points with the objective of reducing off-line planning time.

Without considering the orientation of the EE, the base position and the working area for the mobile manipulator are determined using the BWLS. Compared to other methods, the proposed algorithm is beneficial when the planning problem has six dimensions, ensuring the reachability and stability of the EE.

The algorithm needs no manual configuration, and its performance is investigated for typical surfaces in practical applications.

The purpose of this paper is to provide a review of the interdisciplinary research field, autonomous industrial mobile manipulation (AIMM), with an emphasis on physical implementations and applications.

Following an introduction to AIMM, this paper investigates the missing links and gaps between the research and developments efforts and the real‐world application requirements, in order to bring the AIMM technology from laboratories to manufacturing environments. The investigation is based on 12 general application requirements for robotics: sustainability, configuration, adaptation, autonomy, positioning, manipulation and grasping, robot‐robot interaction, human‐robot interaction, process quality, dependability, and physical properties.

The concise yet comprehensive review provides both researchers (academia) and practitioners (industry) with a quick and gentle overview of AIMM. Furthermore, the paper identifies key open issues and promising research directions to realize real‐world integration and maturation of the AIMM technology.

This paper reviews the interdisciplinary research field, autonomous industrial mobile manipulation (AIMM).

Humanitarian de‐mining tasks require the use of specific detecting sets to detect landmines. These sets are normally based on a one‐point sensor, which must be moved over the infested terrain by a combination of a scanning manipulator and a mobile platform. The purpose of this paper is to present the development of the sensor head and the scanning manipulator.

The manipulator needs sensors in order to negotiate ground irregularities and detect obstacles in the path of the mine‐detecting set. All of the sensors must be integrated into a sensor head that is in charge of both detecting land mines and providing overall sensor functions for the mobile platform's steering controller.

The sensor head is based on a commercial mine‐detecting set and a ground‐tracking set based on a network of range sensors tailor‐made for this purpose; the scanning manipulator is based on a mechanism with five degrees of freedom.

The design assessment and some experiments are reported.

The paper aims to present a tracked robot comprised of several biochemical sampling instruments and a universal control architecture. In addition, a dynamic motion planning strategy and autonomous modules in sampling tasks are designed and illustrated at length.

Several sampling instruments with position tolerance and sealing property are specifically developed, and a robotic operation system (ROS)-based universal control architecture is established. Then, based on the system, two typical problems in sampling tasks, i.e. arm motion planning in unknown environment and autonomous modules, are discussed, implemented and tested. Inspired by the idea of Gaussian process classification (GPC) and Gaussian process (GP) information entropy, three-dimensional (3D) geometric modeling and arm obstacle avoidance strategy are implemented and proven successfully. Moreover, autonomous modules during sampling process are discussed and realized.

Smooth implementations of the two experiments justify the validity and extensibility of the robot control scheme. Furthermore, the former experiment proves the efficiency of arm obstacle avoidance strategy, while the later one demonstrates the time reduction and accuracy improvement in sampling tasks as the autonomous actions.

The proposed control architecture can be applied to more mobile and industrial robots for its feasible and extensible scheme, and the utility function in arm path planning strategy can also be utilized for other information-driven exploration tasks.

Several specific biochemical sampling instruments are presented in detail, while ROS and Moveit! are integrated into the system scheme, making the robot extensible, achievable and real-time. Based on the control scheme, an information-driven path planning algorithm and automation in sampling tasks are conceived and implemented.