Multi-level manipulators are those mechanisms in which two or more levels, that define the main chain of the manipulator, are joined in parallel to each other. Besides, each level is linked to the base in parallel by some limbs. Based on the idea of multi-level manipulators and using the concept of plain leg surfaces, the synthesis of partially decoupled manipulators with five degrees of freedom is presented. Among the different possibilities that exist to design the main chain of the manipulator, one is selected and the different manipulators that can be obtained from this option are analyzed. The concept of control distribution per level is presented and compared with the distribution of degrees of freedom per level. Finally, each of the proposed manipulators is studied and those which decouple the rotations are chosen.

In the original version of this paper the email of the corresponding author, Reza Mahboobi Esfanjani, was given incorrectly. The correct email is [email protected].

The publisher apologises for this error.

A work cell with multiple robots increases manufacturing flexibility and productivity. Robots in the work cell equipped in a sparse area usually share a motion path, resources and workspace. In this paper, an embedded Markov chain model for a multi-robot system that has a common workspace is constructed on the basis of the concept of a multi-processor system. With the presented model, we measure the performance of error recovery schemes under different workloads and analyze the sensitivity of the execution time with respect to the robot speed. We verify the presented model with an experimental multi-robot work cell. This study is useful in evaluating the performance of a robotic work cell and presents a guide for designing a complex work cell.

This study is concerned with the tracking control problem for nonlinear uncertain robotic systems in the presence of unknown actuator nonlinearities. A novel adaptive sliding controller is designed based on a robust disturbance observer without any prior knowledge of actuator nonlinearities and system dynamics. The proposed control strategy can guarantee that the tracking error eventually converges to an arbitrarily small neighborhood of zero. Simulation results are included to demonstrate the effectiveness and superiority of the proposed strategy.

This paper investigates the contributions of each joint in perturbed balance by employing multiple balance strategies and exploring gain scheduling. Hybrid controllers are developed for sagittal standing in response to constant pushes, and a hypothesis is then investigated that postural feedback gains in standing balance should change with perturbation size via an optimization approach. Related research indicates the roles of each joint: the ankles apply torque to the ground, the hips and/or arms generate horizontal ground forces, and the knees and hips squat. To investigate it from an optimization point of view, this paper uses a horizontal push of a given size, direction, and location as a perturbation, and optimizes controllers for different push sizes, directions, and locations. It applies to the ankle, hip, squat, and arm swinging strategies in standing balance. By comparing the capability of handling disturbances and investigating the feedback gains of each strategy, this paper quantitatively analyzes the contributions of each joint to perturbed balance. We believe this work is also instructive to study the progressive behavioral changes as the model gets more and more complex.

The assist-as-needed (AAN) controller is effective in robot-assisted rehabilitation. However, variations of the engagement of subjects with fixed controller often lead to unsatisfying results. Therefore, adaptive AAN that adjusts control parameters based on individualized engagement is essential to enhance the training effect further. Nevertheless, current approaches mainly focus on the within-trial real-time engagement estimation, and the presence of measurement noise may cause improper evaluation of engagement. In addition, most studies on human-in-loop optimization strategies modulate the controller by greedy strategies, which are prone to fall into local optima. These shortcomings in previous studies could significantly limit the efficacy of AAN. This paper proposes an adaptive AAN to promote engagement by providing subjects with a subject-adaptive assistance level based on trial-wise engagement estimation and performance. Firstly, the engagement is estimated from energy information, which assesses the work done by the subject during a full trial to reduce the influence of measurement outliers. Secondly, the AAN controller is adapted by Bayesian optimization (BO) to maximize the subject’s performance according to historical trial-wise performance. The BO algorithm is good at dealing with noisy signals within limited steps. Experiments with ten healthy subjects resulted in a decrease of 34.59$\%$ in their average trajectory error, accompanied by a reduction of 9.71$\%$ in their energy consumption, thus verifying the superiority of the proposed method to prior work. These results suggest that the proposed method could potentially improve the effect of upper limb rehabilitation.

The traditional identification methods of industrial robots are based on Inverse Dynamic Identification Model (IDIM). Based on the model, input torque, motor-side and link-side motion data are necessary when joint flexibilities are considered. However, it is often unavailable or expensive to general robots which are not equipped with link-side sensors. To solve the problem, a novel dynamic parameter identification method, which only employ input torque and motor-side motion data, is proposed in this report. Based on motor-side dynamics, link-side dynamics are modified as high-order nonlinear functions of input torque and motor-side motion. Then, through different trajectory-load groups and high-order observers, the nonlinear equations can be solved, and dynamic parameters can be estimated with short operation time. The selection rules of the trajectory-load groups are then discussed based on simulation results, so as to promote estimation results. Finally, experiments are conducted to verify the proposed method and exhibit the selection rules of observer gains. As shown in the report, except viscous friction parameters, identification difference between the IDM-based methods and the proposed one is less than 9%.

In this paper, the kinematic workspace characteristics of a crab-like legged vehicle are investigated using a 2-D model. The alternative kinematic configurations and their corresponding workspace constraints are discussed, and the vehicle configuration of most interest identified. It is shown that, for constant vehicle body attitude, only two parameters affect the kinematic workspace, foot overlap and thigh length. Analytical methods for calculating the workspace characteristics are presented and, using these methods, the effects of the design geometry on the kinematic workspace are investigated.

A negative pressure wall-climbing robot is a special robot for climbing vertical walls, which is widely used in construction, petrochemicals, nuclear energy, shipbuilding, and other industries. The mobility and adhesion of the wheel-track wall-climbing robot with steering-straight mode are significantly decreased on the cylindrical wall, especially during steering. The reason is that the suction chamber may separate from the wall and the required driving force for movement increases, during steering. In this paper, a negative pressure wall-climbing robot with omnidirectional movement mode is developed. By introducing a compliant adjusting suction mechanism and omni-belt wheels, an omnidirectional movement mode is formed instead of the steering-straight mode, and the performances of adhesion and mobility are improved. We establish the safety adhesion model for the robot on a cylindrical wall and obtain the safety adhesion forces. We designed and manufactured an experimental prototype based on the analysis. Experiments showed that the robot has the ability of full maneuverability in cylindrical walls.

Today, robots can be found helping humans with their daily tasks. Some tasks require the robot to visit a set of locations in the environment efficiently, like in the Traveling Salesman Problem. As indoor environments are maze-like areas, feasible paths connecting locations must be computed beforehand, so they can be combined during the scheduling, which can be impracticable for real-time applications. This work presents an on-line Route Scheduling supported by a Fast Path Planning Method able to adjust pre-built paths. Experiments were carried out with virtual and real robots to evaluate time and quality of tours.

A mathematical model is proposed for a revolute joint mechanism with an n-degree of freedom (DOF). The matrix approach is used for finding the relation between two consecutive links to determine desired link parameters such as position, velocity and acceleration using the forward kinematic approach. The matrix approach was confirmed for a proposed 10 DOF revolute type (R-type) human upper limb model with servo motors at each joint. Two DOFs are considered each at shoulder, elbow and wrist joint, followed by four DOF for the fingers. Two DOFs were considered for metacarpophalangeal (mcp) and one DOF each for proximal interphalangeal (pip) and distal interphalangeal (dip) joints. MATLAB script function was used to evaluate the mathematical model for determining kinematic parameters for all the proposed human upper limb model joints. The simplified method for kinematic analysis proposed in this paper will further simplify the dynamic modeling of any mechanism for determining joint torques and hence, easy to design control system for joint movements.

In November 2012, the eleventh Workshop on Distributed Autonomous Robotic Systems (DARS 2012) was held in Baltimore, Maryland on the campus of Johns Hopkins University. Previous DARS workshops were held in 1992, 1994, and 1996 in Riken, Wako, Japan; Karlsruhe, Germany (1998); Knoxville, Tennessee, USA (2000); Fukuoka, Japan (2002); Toulouse, France (2004); Minneapolis, Minnesota, USA (2006); Tsukuba, Ibaraki, Japan (2008); and Lausanne, Switzerland (2010).

This paper presents the control system of a chess-playing robot developed by the Research Institute of Robots at the Shanghai Jiaotong University. Thanks to the Windows NT operation system and the RTX (Real-Time eXtension), the whole system can achieve good real-time performance. The control system, which is supported by a standard PC hardware platform and a modularized structure of system software, is open-ended and easily expansible.

Operation of tools has long been studied in robotics. Although appropriate hold of the tool by robots is the base of successful tool operation, it is not with ease especially for tools with complicated shape. In this paper, an assist system for a four-limbed robot is proposed for remote operation of reaching and grasping electric drills using deep reinforcement learning. Through comparative evaluation experiments, the increase of success rate for reaching and grasping is verified and the decrease in both physical and mental workload of the operator is also validated by the index of NASA-TLX.

In this paper we present an experimental comparison between different decentralized controllers for industrial robot manipulators. In spite of the fact that robot manipulators are highly nonlinear and coupled multivariable systems, decentralized controllers have been widely adopted in industrial environments for trajectory tracking, because of their simplicity and their fault tolerance feature. Different kinds of controllers have been investigated: the typical PID-based controller, a nonlinear three-term controller devised by Tarokh, and a variable structure control law, a discontinuous integral control, both in the continuous and in the discretised version. Allowances are made for the tuning of the parameters in order to obtain low tracking errors. Results show that better performances are obtained by controllers other than PID, especially by the discontinuous integral control, so that they appear to be particularly appropriate for the use in industrial robots.

To set the sleeping mode for the Yutu-2 rover, a visual pose prediction algorithm including terrain reconstruction and pose estimation was first studied. The terrain reconstruction precision is affected by using only the stereo navigation camera (Navcam) images and the rotation angles of the mast. However, the hazard camera (Hazcam) pose is fixed, and an image network was constructed by linking all of the Navcam and Hazcam stereoimages. Then, the Navcam pose was refined based on a multiview block bundle adjustment. The experimental results show that the mean absolute errors of the check points in the proposed algorithm were 10.4 mm over the range of $\boldsymbol{L}$ from 2.0 to 6.1 m, and the proposed algorithm achieved good prediction results for the rover pose (the average differences of the values of the pitch angle and the roll angle were −0.19 degrees and 0.29 degrees, respectively). Under the support of the proposed algorithm, engineers have completed the remote setting of the sleeping mode for Yutu-2 successfully in the Chang’e-4 mission operations.

An eight-legged walking machine with purely mechanical control of leg motion has been built, based on an old design by Space-General Corporation. The paper describes its mechanical design, and experiments with it on a variety of surfaces.