Upper limb redundancy resolution under gravitational loading conditions: Arm postural stability index based on dynamic manipulability analysis

Abstract

Abstract-Resistance training may be considered as one promising approach for improving the motor capabilities of post-stroke patients. A successful introduction of this depends on the proper resolution of human arm redundancy under gravitational loading. The spatially heterogeneous changes of the human arm swivel angle (which represents the upper limb redundancy) are studied under different loading conditions, the effects of which are incorporated into a modified dynamic manipulability ellipsoid model. A new scalar index describing the arm postural stability (APSI) is then proposed. As part of the experimental protocol, ten (10) healthy subjects performed multiple reaching tasks with different weights mounted on the forearm. Kinematic data was collected via a ten-camera motion capture system and the corresponding APSI was calculated for each task. APSI is found to have a strong linear correlation with the swivel angle under loading conditions. Furthermore, the data suggest that the swivel angle may serve as an indicator of arm postural stability and task difficulty. The results of additional experiments conducted with three (3) subjects indicate that the external loads could deteriorate the arm's control performance in tasks like line tracing (root mean square deviation from straight lines). These findings may be applicable to robot-based (exoskeleton) resistance therapy, assist-as-needed gravity compensation, and human-like motion control of humanoid robotic systems.

Asymmetric Dual Arm Approach For Post Stroke Recovery Of Motor Functions Utilizing The EXO-UL8 Exoskeleton System: A Pilot Study

“…With the external assistance, even in dual-arm task mode the subject could reach more targets which need coordinated shoulder flexion and elbow extension. This matches the results found in our previous research using a manipulability model [14]. Another interesting observation (not plotted) is that without additional instructions, the subject tended to rest his right arm and wait for assistance instead of trying to initiate the reach (rest period is not included in time calculation).…”

Admittance Control Scheme Comparison of EXO-UL8: A Dual-Arm Exoskeleton Robotic System

“…In addition, asymmetric bilateral training using an interactive virtual reality environment was also developed [16]. In either of the training modes and regardless of the motor control functionality level of the human operator, the exoskeleton arm needs to have high "transparency", which means that it is sensitive enough, in a heuristic way, to not exert unwanted resistance on the human arm (otherwise the human's movement, e.g., redundancy resolution, would be affected [17]), as well as be robust enough to prevent instability in all configurations. A trade-off between sensitivity and stability exists, and thus a dynamic equilibrium is needed.…”

“…9. To quantify the error, let it be defined as the root-mean-square (RMS) of the minimum distance of the actual trajectory to the reference trajectory summed over all segments of the reference trajectory, similar to the approach presented in [17]:…”

Assistive Arm-Exoskeleton Control Based on Human Muscular Manipulability

“…In a study (Petrič et al, 2016), we proposed a control approach that compensates the anisotropic property of the kinematic manipulability related to the human arm. Similarly, in Shen et al (2017), the control approach was improved by incorporating endpoint loading conditions into the modified manipulability models. In Kim et al (2010), the dynamic manipulability was used to generate an energy efficient gait pattern.…”

“…In Kim et al (2010), the dynamic manipulability was used to generate an energy efficient gait pattern. However, these control methods (Kim et al, 2010; Petrič et al, 2016; Shen et al, 2017) were only based on human limb kinematics without considering also biomechanical specifics of human arms, i.e., muscles. In contrast to typical robotic actuators with gears and motors, the human joints are actuated by sets of antagonistically coupled muscles.…”