Modeling,observation, and control of hysteresis torsion in elastic robot joints

Hysteresis torsion in elastic robot joints occurs as a coupled nonlinearity due to internal friction, backlash, and nonlinear stiffness, which are coactive inside of mechanical transmission assemblies. The nonlinear joint torsion leads to hysteresis lost motion and can provoke control errors in relation to the joint output at both trajectories tracking and positioning. In this paper, a novel modeling approach for describing the nonlinear input–output behavior of elastic robot joints is proposed together with the observation and control method, which aim to compensate for the relative joint torsion without load sensing. The proposed modeling approach includes the recently developed 2SEP dynamic friction model and Bouc–Wen-like hysteresis model, which is originated from structural mechanics, both arranged according to the assumed torque transmitting structure. The proposed method is evaluated with experiments using the laboratory setup which emulates a single rotary joint under impact of nonlinear elasticities, friction, and gravity.     

An identification procedure for rate-dependency of friction in robotic joints with limited motion ranges

This paper proposes a procedure for identifying rate-dependent friction of robotic manipulators of which the motion is limited due to the configuration or the environment. The procedure is characterized by the following three features: (i) the rate dependency is represented by line sections connecting sampled velocity-force pairs, (ii) the robot is position-controlled to track desired trajectories that are some cycles of sinusoidal motion with different frequencies, and (iii) each velocity-force pair is sampled from one cycle of the motion with subtracting the effects of the gravity and the inertia. The procedure was validated with a six-axis industrial robotic manipulator YASKAWA MOTOMAN-HP3J, of which the joints are equipped with harmonic-drive transmissions of the reduction ratios of 81.5–224. The experimental results show that the identification is achieved with a sufficient accuracy with the 20 degrees of motion of each joint. In addition, the results were utilized for friction compensation, successfully reducing the effect of the friction by 60–80%.     

State estimation on flexible robots using accelerometers and angular rate sensors

Fast moving robots with elastic joints and/or elastic links require vibration suppression to achieve accuracy. We introduce a two-degrees-of-freedom control scheme (2DoF) consisting of a feedforward component based on the flatness approach and a classic feedback component, such as a PD motor joint control with additional link position error or joint torque error feedback. Such a control scheme requires knowledge of the full system state. We present an approach for state estimation using angular rate-and acceleration-sensors-mounted on each robot link-that can be used for any kind of elastic robot manipulator. We validate our theory by measuring a very fast trajectory and the influence of an external disturbance on an articulated robot with two flexible links and three joints.     

An experimental study of planar impact of a robot manipulator

An experimental study of planar impact of a robot manipulator with a stationary rigid surface is presented. Using the data collected from a series of experiments, this paper investigates the post-impact behavior for different pre-impact conditions such as the configuration of the robot, the angle and the velocity of impact. A better understanding of the post-impact behavior for various pre-impact conditions can lead to improved control strategies during impacts in robotic manipulation. Potential applications of this study include robotic assembly and contact tasks. Further, this study could be seen as a preliminary experimental work on impact of general kinematic chains.     

Inverse Double NARX Fuzzy Modeling for System Identification

In this paper, a novel inverse double nonlinear autoregressive with exogenous input (NARX) fuzzy model is applied to simultaneously model and identify both joints of the prototype two-axis pneumatic artificial muscle (PAM) robot arm's inverse dynamic model. Highly nonlinear features of both joints of the nonlinear manipulator system are identified by the proposed inverse double NARX fuzzy (IDNF) model based on experimental input–output training data. The modified genetic algorithm (GA) optimally generates the appropriate fuzzy if–then rules to perfectly characterize the dynamic features of the two-axis PAM manipulator system. The evaluation of different IDNF models with various ARX model structures will be discussed. For the first time, the nonlinear IDNF model of the two-axis PAM robot arm is investigated. The results show that the nonlinear IDNF model that is trained by GA performs better and has a higher accuracy than the conventional inverse fuzzy model.      

Visual control of robotic manipulator based on neural networks

A control scheme for a robotic manipulator system that uses visual information to position and orient the end-effector is described. The control system directly integrates visual data into the servoing process without subdividing the process into determination of the position and orientation of the workplace and inverse kinematic calculation. The feature of the control scheme is the use of neural networks for the determination of the change in joint angles required in order to achieve the desired position and orientation. The proposed system is able to control the robot so that it can approach the desired position and orientation from arbitrary initial ones. Simulations for a robotic manipulator with six degrees of freedom are described. The validity and the effectiveness of the proposed control scheme are verified by computer simulations     

Global minimum-jerk trajectory planning of robot manipulators

A new approach based on interval analysis is developed to find the global minimum-jerk (MJ) trajectory of a robot manipulator within a joint space scheme using cubic splines. MJ trajectories are desirable for their similarity to human joint movements and for their amenability to path tracking and to limit robot vibrations. This makes them attractive choices for robotic applications, in spite of the fact that the manipulator dynamics are not taken into account. Cubic splines are used in a framework that assures overall continuity of velocities and accelerations in the robot movement. The resulting MJ trajectory planning is shown to be a global constrained minimax optimization problem. This is solved by a newly devised algorithm based on interval analysis and proof of convergence with certainty to an arbitrarily good global solution is provided. The proposed planning method is applied to an example regarding a six-joint manipulator and comparisons with an alternative MJ planner are exposed     

Motion control of three-link brachiation robot by using final-statecontrol with error learning

This paper demonstrates that motion control of a three-link brachiation robot can be performed via final-state control for a linear parameter-varying system with error learning. Since the root joint of the brachiation robot has no drive unit, the control problem is treated as that for an open-chain manipulator possessing nondriven joints in circumstances with gravity. Also, the brachiation robot possesses a nonlinear property in its wide motion area from an initial state to a final desired state. We treat the brachiation robot as a discrete-time linear parameter-varying system by using an extended linearization method and Euler's method. In the experiments, the second and third joints of the brachiation robot are directly driven by AC servo motors installed in each joint. The free vibration of the brachiation robot is coincident with that of the simulation and gives suggestions of the final time to reach a desired state that has to be specified for final-state control. By simulations and experiments, it is demonstrated that the feedforward input obtained by the final-state control method can bring the three-link brachiation robot from an initial state to a desired state in a specific time. Its robustness under a variation of the link mass is also verified     

Development of a Biofeedback Therapeutic-Exercise-Supporting Manipulator

Although much equipment for physical therapy has been developed, equipment to improve the quality of physical therapy is scarce. We propose a robotic biofeedback exercise device that can display human joint torque and muscle force during training without a problematic electromyogram (EMG). The purpose is to increase the therapeutic value by understanding a person's condition during exercise and to provide an incentive to improve performance. The manipulator supports lower limb rehabilitation in the sagittal plane. With its ability to adjust the maximum speed and the time constant, the manipulator provides simultaneous and safe isokinetic exercise for the knee and hip joints. This paper describes the estimation of the human joint torque and muscle force. The display of the joint torque and the muscle force is realized during exercise of the knee joint using the developed manipulator. The estimation of the muscle force from Crowninshield's method and Hase's method generally agrees with the EMG.     

A Fault-Tolerant Manipulator Robot Based on ${{cal H}}_2$, ${{cal H}}_{infty }$, and Mixed ${{cal H}}_2/{{cal H}}_{infty }$ Markovian Controls

This paper develops a Markovian jump model to describe the fault occurrences in a manipulator robot of three joints. This model includes the changes of operation points and the probability that a fault occurs in an actuator. After a fault, the robot works as a manipulator with free joints. Based on the developed model, a comparative study among three Markovian controllers, H₂, H_infin, and mixed H₂/H_infin is presented, applied in an actual manipulator robot subject to one and two consecutive faults.     

Effects of joint dynamics on the dynamic manipulability of geared robot manipulators

Information obtained through the analysis of dynamic manipulability can be useful in the design and control of robotic manipulators. In this paper, we investigate the influences of the joint drive systems on the dynamic manipulability, with special attention to the effects of gear reduction ratios. In order to reflect this influence, a new form of the dynamic manipulability ellipsoid is developed, which includes the gear reduction ratios, and a modified formula is derived for the dynamic manipulability measure. This formula indicates that joint drive dynamics can significantly influence the dynamic manipulability of a robot. The most important effect is that changing the gear reduction ratios modifies the nature of the dynamic manipulability distribution throughout the workspace of the manipulator. Examples of a two-link planar arm and a PUMA 560 robot are presented to illustrate the results.     

Rapid Prototyping of a Model-Based Control With Friction Compensation for a Direct-Drive Robot

The development and application of the most recent model-based control schemes for robots require the investigation and solution of problems concerning various aspects, from the real-time (RT) simulation and control issues, to the necessity of determining a robot model suitable for control, and of experimentally testing the control performances. In this paper, different aspects are investigated for a planar, two-link direct-drive manipulator, with particular attention to the joint friction compensation within the control loop. A real-time architecture, based on the RT-Lab software by Opal-RT, is developed and used to carry out all the design phases, from the identification of the robot model, including joint friction torques, to the application of the inverse dynamics control schemes, with different solutions for the robot dynamics and friction compensation. The performances of such schemes are investigated executing various trajectories, suitable to check the effectiveness of the friction compensation.     

A mechatronic approach for robotic deburring

This paper deals with the implementation of a mechatronic methodology for the robotic deburring of planar workpieces with an unknown shape performed by an industrial manipulator. The approach is based on the use of a hybrid force/velocity control law and on a correlated suitable design of the deburring tool. Experimental results, obtained with a two degrees-of-freedom SCARA industrial robot manipulator, show the effectiveness of the method.     

Development of 3D Scanning System for Robotic Plasma Processing of Medical Products with Complex Geometries

This paper describes the development of an intelligent automated control system of a robot manipulator for plasma treatment of medical implants with complex shapes. The two-layer coatings from the Ti wire and hydroxyapatite powders are applied on the surface of Ti medical implants by microplasma spraying to increase the biocompatibility of implants. The coating process requires precise control of a number of parameters, particularly the plasma spray distance and plasma jet traverse velocity. Thus, the development of the robotic plasma surface treatment involves automated path planning. The key idea of the proposed intelligent automatic control system is the use of data of preliminary three-dimensional (3D) scanning of the processed implant by the robot manipulator. The segmentation algorithm of the point cloud from laser scanning of the surface is developed. This methodology is suitable for robotic 3D scanning systems with both non-contact laser distance sensors and video cameras, used in additive manufacturing and medicine.     

A disturbance observer for robot manipulators with application toelectronic components assembly

The success of robotic assembly for odd form electronic components depends heavily on the ability to monitor and control the insertion force. For this purpose, a joint disturbance observer for robot manipulators is presented to estimate the reaction force due to component misinsertion. The derivation is presented in the time domain for a multi-degree-of-freedom manipulator without making the SISO linear system assumption. Based on the internal model control (IMC) concept, this observer is simple in structure. It is in the form of an integral controller and only requires measurements of joint position and velocity signals to be implemented. The observed perturbation signal includes the effects of model uncertainties as well as reaction force against the environment. Careful trajectory planning and detailed friction modeling are performed to enhance accuracy of the deduced reaction force. Experiments conducted on a SCARA robot for the insertion of PCB components demonstrated the performance of the proposed disturbance observer     

A nonlinear momentum observer for sensorless robot collision detection under model uncertainties

Collision detection methods could reduce collision forces and improve safety during physical human-robot interaction without additional sensing devices. However, current collision detection methods result in an unavoidable trade-off between sensitivity to collisions, peaking value reduction near the initial time, and immunity to measurement noise. In this paper, a novel nonlinear extended state momentum observer (NESMO) is proposed for detecting collisions between a robot body and human under model uncertainties based on only position and current measurements. The collision detection method is divided into three steps. The first step is to identify the robot dynamic model. Then, we can deduce the generalized momentum-based state-space equations from the identified base dynamic parameters. The second step is to construct a NESMO. Benefiting from the fractional power function and the time-varying damping ratio, the NESMO achieves the required monitoring bandwidth with noise immunity. The last step is to design a novel time-varying threshold (TVT) to distinguish the collision signal from the estimated lumped disturbance. As with the dynamic model parameters, the coefficients of TVT could be obtained by offline identification. Combined with NESMO, the method can provide timely and reliable collision detection and estimation under model uncertainties. Simulation and experimental results obtained using a 6-DOF robot manipulator illustrate the effectiveness of the proposed method.     

Control of Adept One SCARA robot using neural networks

This paper presents an enhanced feedback error learning control (EFELC) strategy for an n-degree-of-freedom robotic manipulator. It covers the design and simulation study of the neural network-based controller for the manipulator with a view of tracking a predetermined trajectory of motion in the joint space. An industrial robotic manipulator, the Adept One Robot, was used to evaluate the effectiveness of the proposed scheme. The Adept One Robot was simulated as a three-axis manipulator with the dynamics of the tool (fourth link) neglected and the mass of the load incorporated into the mass of the third link. For simplicity, only the first two joints of the manipulator were considered in the simulation study. The overall performance of the control system under different conditions, namely, trajectory tracking, variations in trajectory and different initial weight values were studied and comparison made with the existing feedback error learning control strategy. The enhanced version was shown to outperform the existing method     

Adaptive robot control using neural networks

This paper studies the trajectory tracking problem to control the nonlinear dynamic model of a robot using neural networks. These controllers are based on learning from input-output measurements and not on parametric-model-based dynamics. Multilayer recurrent networks are used to estimate the dynamics of the system and the inverse dynamic model. The training is achieved using the backpropagation method. The minimization of the quadratic error is computed by a variable step gradient method. Another multilayer recurrent neural network is added to estimate the joint accelerations. The control process is applied to a two degree-of-freedom (DOF) SCARA robot using a DSP-based controller. Experimental results show the effectiveness of this approach. The tracking trajectory errors are very small and torques expected at manipulator joints are free of chattering.<>     

Interaction control of an industrial manipulator using LPV techniques

This paper presents the design and implementation of a hybrid force/motion control scheme on a six-degrees-of-freedom robotic manipulator employing a gain-scheduled linear parameter-varying (LPV) controller. A nonlinear dynamic model of the manipulator is obtained and the unknown parameters are estimated. The manipulator is decomposed into an inner and a wrist submodel, and a practical way is proposed to investigate the coupling between them. The motion control part of the hybrid controller which is the main focus of this paper is formed by a combination of an LPV controller and a model-based inverse dynamics controller for the inner submodel and the wrist joints, respectively. A quasi-LPV model with a reduced number of scheduling parameters is derived for the inner submodel, and a polytopic LPV gain-scheduled controller is synthesized in a two-degrees-of-freedom structure including feedback and feedforward parts, which is augmented by a friction compensation term. A PD controller with a feedforward path is designed to control the interaction force. The proposed hybrid force/motion scheme is implemented on the 6-DOF CRS A465 robotic manipulator to perform a writing task. Comparison of the results with those of a hybrid force/motion controller with a plain model-based inverse dynamics motion control and the same force control shows that the proposed controller improves the position tracking performance significantly.     

Kinematics modeling and performance optimization of a kinematic-mechanics coupled continuum manipulator

An optimization design of a notched continuum manipulator is proposed in this paper, which is based on our defined performance evaluation indices. By experimental testing the mechanics-based forward kinematics and the curve-fitting-based inverse kinematics, our proposed kinematic-mechanics coupled continuum manipulator is equivalent to a robot that features discrete joints and rigid links. Based on the mechanics and the kinematics models, two kinematic performance evaluation indices are developed to comprehensively describe the performance distribution characteristics in the workspace. Firstly, the global stiffness performance index of the proposed manipulator is established based on the 3-dimensional 2-node Timoshenko beam element. Secondly, the equivalent Jacobian matrix, a linear mapping from the joint space to the task space of the continuum manipulator, is analyzed, and equivalent dexterity performance index is determined. Finally, the optimal design for our proposed continuum manipulator considering the stiffness performance and kinematics performance simultaneously is implemented as a case study utilizing the NAGA-II genetic algorithm. The results demonstrate that the optimized geometric parameters can improve the kinematic performance evaluation indices simultaneously, and our proposed kinematic performance evaluation indices can be applied and extended to the kinematic evaluation and the kinematic optimization of this kind of notched continuum manipulator.