Human Arm Motion Analysis Based on the Inverse Optimization Approach

In this article, we explore human motion skills in the dual-arm manipulation tasks that include contact with equipment with the final aim to generate human-like humanoid motion. Human motion is analyzed using the optimization approaches starting with the assumption that human motion is optimal. A combination of commonly used optimization criteria in the joint space with the weight coefficients is considered: minimization of kinetic energy, minimization of joint velocities, minimization of the distance between the current and ergonomic positions, and maximization of manipulability. The contribution of each criterion for seven different dual-arm manipulation tasks to provide the most accurate imitation of the human motion is given via suggested inverse optimization approach calculating values of weight coefficients. The effects on actors’ body characteristics and the characteristics of the environment (involved equipment) on the choice of criterion functions are additionally analyzed. The optimal combination of weight coefficients calculated by the inverse optimization approach is used in our inverse kinematics algorithm to transfer human motion skills to the motion of the humanoid robots. The results show that the optimal combination of weight coefficients is able to generate human-like humanoid motions rather than individual one of the considered criterion functions. The recorded human motion and the motion of the humanoid robot ROMEO, obtained with the strategy used by human and defined by our inverse optimal control approach, for the tasks “opening/closing a drawer” are assessed visually and quantitatively.

Manual GTAW is commonly used in industry especially for applications which require high weld quality. Unfortunately, skills needed for critical welding operations typically require a long time to develop. In this paper a control system was developed to assist the welder in regulating his/her arm movement (i.e., the welding speed) in real-time. Training experiments are conducted by a human welder, whose movement speed (system output) is accurately tracked by a Leap motion sensor. Visual commands (arrows with both direction and amplitude) are displayed on a monitor, which is used as system input. Modeling trials suggest that Moving Average (MA) models are not sufficient in capturing the human movement dynamics. Auto-Regressive Moving Average (ARMA) models are then identified using Least Squares (LS), and F-test is used to select the model order. Pole placement controller is proposed to control the human welder arm movement. Controller simulation shows that the proposed controller is able to drive the system to desired speed with no static error and acceptable regulating speed. To further demonstrate the effectiveness of the proposed controller, tracking experiments are conducted by a human welder to track various set-points. It is observed that with the proposed pole placement controller, the welder can track the desired speed with acceptable accuracy. The proposed welding speed assistance system may be beneficial to develop the next generation intelligent welding machines.

Manipulator Control Using Human Arm Imitation

Event Abstract Back to Event Motor equivalence reveals coarticulation in arm movement sequences involving obstacle avoidance Eva Nowak1*, Britta Grimme1, Hendrik Reimann1 and Gregor Schöner1 1 Ruhr-Universität Bochum, Institut für Neuroinformatik, Germany We study movement sequences involving obstacle avoidance in naturalistic 3D arm movements in humans looking for coarticulation. We use the concept of the uncontrolled manifold (Scholz & Schoener, 1999) to uncover motor equivalent solutions when movements in one segment are preceded by different movements in the preceding segment. In the experiment, participants moved an object with their hand along a horizontal path from one target to the next while avoiding obstacles placed between the targets. Each of four segments began with tapping the object on a target, then moving over an obstacle and tapping on the next target, which started the following segment. The obstacles varied in height, higher obstacles being placed in the second and third segments of the movement. We looked for coarticulation, that is, for influences of these higher obstacles on the first and last segments. Avoiding lower obstacles immediately after avoiding higher obstacles resulted in adaption of the spatial path of the transported object (post-coarticulation). This adaption increased with increasing height of the preceding obstacle. At the joint level we decomposed the variance of joint configurations into components that leave the spatial position of the transported object invariant (UCM) and the orthogonal complement (ORT). We found higher variance within the UCM than within ORT across the entire movement sequence. Coarticulation was addressed at the joint level by decomposing the difference between joint configurations in different conditions into a UCM and an ORT component. The UCM component is motor equivalent because it leaves the task variable invariant. The difference between joint configurations in a given segment, when different obstacle conditions were experienced in the preceding segment, lay largely in the UCM. Such motor equivalence is evidence for coarticulation at the level of the joint angles. Anticipatory coarticulation was not found consistently across all participants. Acknowledgements The authors acknowledge support from the German Federal Ministry of Education and Research within the National Network Computational Neuroscience - Bernstein Fokus: "Learning behavioral models: From human experiment to technical assistance", grant FKZ 01GQ0951. References Scholz JP, Schoner G. (1999) The uncontrolled manifold concept: identifying control variables for a functional task. Exp Brain Res 126: 289–306 Keywords: arm movement, coarticulation, motor equivalence, obstacle avoidance, Uncontrolled Manifold Conference: Bernstein Conference 2012, Munich, Germany, 12 Sep - 14 Sep, 2012. Presentation Type: Poster Topic: Motor control, movement, navigation Citation: Nowak E, Grimme B, Reimann H and Schöner G (2012). Motor equivalence reveals coarticulation in arm movement sequences involving obstacle avoidance. Front. Comput. Neurosci. Conference Abstract: Bernstein Conference 2012. doi: 10.3389/conf.fncom.2012.55.00232 Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters. The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated. Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed. For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions. Received: 11 May 2012; Published Online: 12 Sep 2012. * Correspondence: Mrs. Eva Nowak, Ruhr-Universität Bochum, Institut für Neuroinformatik, Bochum, 44780, Germany, eva.nowak@ini.rub.de Login Required This action requires you to be registered with Frontiers and logged in. To register or login click here. Abstract Info Abstract The Authors in Frontiers Eva Nowak Britta Grimme Hendrik Reimann Gregor Schöner Google Eva Nowak Britta Grimme Hendrik Reimann Gregor Schöner Google Scholar Eva Nowak Britta Grimme Hendrik Reimann Gregor Schöner PubMed Eva Nowak Britta Grimme Hendrik Reimann Gregor Schöner Related Article in Frontiers Google Scholar PubMed Abstract Close Back to top Javascript is disabled. Please enable Javascript in your browser settings in order to see all the content on this page.

During the communication and interaction with a human using motions or gestures, a humanoid robot needs to not only look like a human but also behavior like a human to avoid confusions in the communication and interaction. Among human-like behaviors, arm motions of the humanoid robot are essential for the communication with people through motions. In this work, a mathematical representation for characterizing human arm motions is first proposed. The human arm motions are characterized by the elbow elevation angle that is determined using the position and orientation of human hands. That representation is mathematically obtained using an approximation tool, response surface method (RSM). Then, a method to generate human-like arm motions in real time using the proposed representation is presented. The proposed method was evaluated to generate human-like arm motions when the humanoid robot was asked to move its arms from a point to another point including the rotation of hand. An example motion was performed using the KIST humanoid robot, MAHRU

The aim of the research was to determine how the reaction time and the movement speed depend on the complexity of the task. The research was carried out in the Laboratory of Human Motor Control at the Lithuanian Academy of Physical Education (LAPE) applying the analyzer of dynamic parameters of human leg and arm movement (DPA-1; Patent No. 5251; 2005 08 25), which is used for the qualitative estimation of the dynamic parameters of one arm and leg target movement, two arms and legs coordinated and independent target movements, when the resistance power and target are coded with different programmable parameters. We registered the reaction time (RT) and the movement speed (Vmax) performing simple tasks of reaction and speed and a complicated task of accuracy. Research results indicated that performing a complicated task the reaction is slower, and the maximal movement speed is lower than performing a simple task. However, it does not mean that movement speed will be higher when the reaction is faster performing a simple task. The data obtained confirmed Hick’s law proposing that reaction time is directly proportional to the complexity of the task because performing the tasks of different levels of complexity the reaction time values of the right arm were statistically significantly different (p < 0.001). After performing the analysis of variation coefficients we established that the highest coefficient of variation was received from the indices of movement speed performing a speed task (23%), and the lowest — of reaction time performing a reaction task (10%). The obtained results confirm other authors’ suggestion that performing a complicated task the reaction time is a more steadily controlled index than maximal movement speed. A strong correlation was determined between the reaction time values performing the tasks of reaction and accuracy, but there was no statistical link between the maximal speed values performing the tasks of speed and accuracy. This indicates that if the movement speed is high performing a simple task, it does not mean that it will be high performing a complicated task. Conclusions: 1) performing a complicated task reaction time is longer, and maximal movement speed is lower than performing a simple task; 2) the complexity of the task more impacts the dispersion of results of the movement speed than of the reaction time; 3) high speed performing a simple task does not indicate that it will be high performing a complicated task.Keywords: reaction time, movement speed, the complexity of the task.

This paper describes a simple computational model of joint torque and impedance in human arm movements that can be used to simulate three-dimensional movements of the (redundant) arm or leg and to design the control of robots and human-machine interfaces. This model, based on recent physiological findings, assumes that (1) the central nervous system learns the force and impedance to perform a task successfully in a given stable or unstable dynamic environment and (2) stiffness is linearly related to the magnitude of the joint torque and increased to compensate for environment instability. Comparison with existing data shows that this simple model is able to predict impedance geometry well.

A small-gain approach to inverse optimal adaptive control of nonlinear systems with unmodeled dynamics

A common hypothesis in human motor control is that human movement is generated by optimizing with respect to a certain criterion and is task dependent. In this paper, a method to segment human movement by detecting changes to the optimization criterion being used via inverse optimal control is proposed. The control strategy employed by the motor system is hypothesized to be a weighted sum of basis cost functions, with the basis weights changing with changes to the motion objective(s). Continuous time series data of movement is processed using a sliding fixed width window, estimating the basis weights of each cost function for each window by minimizing the Karush-Kuhn-Tucker optimality conditions. The quality of the cost function recovery is verified by evaluating the residual. The successfully estimated basis weights are averaged together to create a set of time varying basis weights that describe the changing control strategy of the motion and can be used to segment the movement with simple thresholds. The proposed algorithm is first demonstrated on simulation data and then demonstrated on a dataset of human subjects performing a series of squatting tasks. The proposed approach reliably identifies the squatting movements, achieving a segmentation accuracy of 84%.

Background: Stroke patients require arm movement exercising for various stimulations in standing position for various stimulations rather than in a sitting position because they require integrated skillful movements, such as stretching, holding, and controlling. Objective: This study was conducted to provide foundational clinical data about lower limb global synkinesis in stroke patients using arm movements in a standing position. Design: Randomized controlled trial. Methods: The subjects were divided into a control group (n = 10) and an experimental group (n = 10), and a pre-test was conducted to evaluate leg global synkinesis (GS) and balance. Intervention method is stretching an arm to hold a ball, repeating supination and pronation of the hand only while maintaining the arm extended as much as possible, repeating shoulder abduction and adduction while holding the pegboard. This was followed by a three-week intervention during which re-measurement was conducted in the same way as was done for the pre-test. Results: The control group showed a significant difference in GS and balance during plantar flexion (p<.05), and the experimental group showed a significant difference in GS and balance during all movements (p<.05, p<.01, respectively). There was a significant difference in GS and balance between the two groups during dorsiflexion (p<.05, p<.01, respectively). Conclusion: The findings demonstrate that human arm movements in a standing position can reduce GS in the affected limb, and balance can be improved by stimulating the surrounding tissues of the affected limb and changing them positively.

The role of sensorimotor (S-M) areas in the specification of kinematic parameters for aiming movements was studied by comparing the performance of six subjects with unilateral stroke to that of matched control subjects. Rapid arm movements were made to one of four targets by rotating the forearm in a short (20 degrees) or long (45 degrees) arc of motion. Thus, the four targets represented two directions (flexion or extension) and two extents (short or long). Subjects with stroke used the arm ipsilateral to the side of the lesion. A timed-response paradigm was used to dissociate response initiation and specification. Subjects initiated movements in concert with the last of four regularly timed tones. A visual cue of the designated target was presented during the preparation interval (400-0 ms) before the last tone. Targets were presented in a fixed sequence (predictable condition) or a random sequence (unpredictable condition). No significant differences in performance were found between stroke and control groups in the predictable condition. In the unpredictable condition, subjects with stroke produced more direction errors and were less accurate in extent than the control subjects. As specification time increased to 400 ms, the frequency of direction errors attenuated less for stroke than for control groups, but the reduction in magnitude of extent errors was similar for the two groups. When specification was minimal (i.e., <100 ms), default responses were distributed equally between directions and clustered around the short extent. Further, wrong direction responses did not converge on the designated extent as specification time increased. This pattern of findings is consistent with a view of parameterization of planning and executing movements, in which direction and extent can be specified in parallel. Our results suggest that ipsilateral S-M areas contribute to the specification of an optimal motor program, particularly when imperative programming of unimanual goal-directed aiming movements is required.

Proposed optimization criteria to explain human multijoint arm motion include smoothness criteria and a minimum-variance criterion based on motor command-dependent noise (signal-dependent noise). It is reported that each of these criteria can effectively reproduce the human trajectory. This paper investigates the possibility that the minimum-variance criterion is included in trajectory planning. First, the relation between the hand trajectory and the final point variance is investigated for motion in the sagittal plane by a numerical experiment. Next, the corresponding relation is investigated by a behavioral experiment. Then, noise is added to the motor command as a simulation, and a behavioral experiment is performed to investigate the relation between the change of the trajectory and the variance. The experimental result suggests the possibility that the minimum-variance criterion is used in trajectory planning. However, it is difficult to account for the human arm motion in the sagittal plane by using only the minimum-variance criterion. The above results seem to indicate that both the smoothness criterion and the minimum-variance criterion based on signal-dependent noise are used in human arm trajectory planning. © 2005 Wiley Periodicals, Inc. Syst Comp Jpn, 36(14): 32–43, 2005; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.20365

The hydraulic robot arm consists of two major parts: first is a human arm movement sensing and transmitting section, and the second is a receiving section and hydraulic robot arm. For sensing, linear potentiometers have been used and the voltage drop across each potentiometer is modeled. The control signals have been generated through CC430F6137 microcontroller and sent wirelessly. On the other end, the receiver captures the respective signals for the human arm movement, decodes it and then sends to the hydraulic robot arm control section. The innovative hydraulic control section has been developed to control the robot arm movement, in which both the suction and pressure mechanism is utilized for the robot arm actions. The developed system has been tested by operating from 200 meter distance by performing the human actions and observed that the robot arm performs a similar action as that of the human arm.

Human arm movements towards visual targets are remarkably reproducible in several tasks and conditions. Various authors have reported that trajectories of unconstrained point-to-point movements are slightly curved, smooth and have bell-shaped velocity profiles. The hand paths of such movements show small - but significant - curvatures throughout the workspace. The cause of these curvatures is still obscure. Traditionally this curvature is explained as the result of an optimisation process or is ascribed to mechanical or dynamic properties of the effector system. Recently, however, it has been suggested that these curvatures are due at least partly, to the visual misperception of straight lines. To evaluate the latter hypothesis, we compared unconstrained, self-paced point-to-point movements that subjects made with their right and left hand. We assume that the visual misperception may depend on the position in the workspace, subject, etc. but not on the hand used to make the movement. Therefore we argue that if curvature is caused by a visual misperception of straight lines, curvatures should be the same for movements made with the left and right hand. Our experiments cast strong doubt on the hypothesis that curvatures are the result of a visual distortion, because curvatures of the left hand trajectories, mirrored in the mid-sagittal plane, are found to be accurately described by trajectories of the right hand. Estimates of the effect of visual distortion on movement curvature show that, if present, this effect is very small compared with other sources that contribute to movement curvature. We found that curvatures depend strongly on the subject and on the direction and distance of the movement. Curvatures do not seem to be caused purely by the dynamic properties of the arm, since curvatures do not change significantly with increasing movement velocity. Therefore, we conclude that curvatures reflect an inherent property of the control of multi-joint arm movements.

In order to explain features of a point-to-point human arm movement, several models based on a minimization principle have been proposed. For example, there are the minimum hand jerk model, the minimum angle jerk model, the minimum commanded torque change model, the minimum-variance model and so on. For some models listed above; the minimum hand jerk model, the minimum angle jerk model and the minimum commanded torque change model, these established methods to obtain an optimal trajectory which satisfies corresponding criterion have already been reported. However it is difficult to obtain the optimal solutions of the minimum-variance model and of other some models. In this study, we propose a novel method to calculate optimal trajectories without depending on the employed criterion by using genetic algorithm. The proposed method has following two procedures; first, the objective trajectory is described by the sum of polynomial equations which satisfy the boundary conditions strictly, and second, the coefficients of the polynomial expressions are determined by using genetic algorithm. In this report, we have confirmed the effectiveness of our method in numerical experiments. First, in the minimum commanded torque change model, we show that the produced trajectory is in beautiful agreement with the optimal solution. Next, in the minimum-variance model, the obtained solution by using proposed method has better evaluated values than one of the quasi-optimal solution that is obtained by the conventional Simplex method