Wrist impairments caused by stroke or long-term immobilization should be given more attention than the rehabilitation of proximal limb segments. In addition, the physiological effect of wrist exoskeletons on the training lacks sufficient research. According to the design requirements, a novel parallel wrist rehabilitation robot (PWRR) with two electric linear actuators and a passive prismatic joint was developed to achieve the training of the flexion/extension (F/E) and radial/ulnar deviation (R/U). The constraint forces caused by the axis misalignment were significantly reduced under the passive and active training by 0.75 ± 0.22 N and 2.50 ± 0.72 N, respectively. The total work of the active training increased with the velocity, whereas it was much smaller and had no obvious linear relationship with the velocity for the passive training. The sequence and degree of muscular activation of targeted muscles were measured and analyzed at three velocities to reveal the effects of the training mode and velocity on the physiological effects. PWRR could achieve the isometric training of wrist muscles at any angle of F/E and R/U validated by the comparative experiments of this device and weights. At each interactive force, the median frequency (MNF) of the EMG power spectrum and the average rectified value (ARV) of its amplitude decreased with increasing training time. The MNF slope increased with interactive thresholds, where the maximum slope was (−29.5±2.42) ×10 ⁻³ %/s at 15 N. Conversely, the ARV slope gradually decreased with the thresholds, where the maximum slope was (−61.0±8.55) ×100 ⁻³ %/s at 5 N.

Background: Trajectory planning is the premise of the control of orthopedic robots, which is directly related to the safety of the human body. However, to date, the trajectory of orthopedic robots has been restricted to lines and spline curves. This limits the flexibility of the robot and leads to unsatisfactory performance. In this paper, a trajectory planning method based on improved RRT* and B-spline curve is proposed in order to improve the control accuracy and flexibility. Method: Firstly, combined with the shortcomings of current trajectory planning methods and bone docking task analysis, the characteristics of the trajectory for orthopedic robot are illustrated, and the problem is described. Secondly, a sampling strategy and an extension strategy are proposed to solve the optimal problem of the RRT* algorithm. Meanwhile, B-spline curve is selected for path smoothing. Thirdly, based on our orthopedic robot, kinematics analysis is introduced briefly, and hypotonic polynomial is used to fit the joint variables. Finally, a comparative study of the improved RRT*, RRT*, and other algorithms are completed, and the feasibility of the robot’s trajectory is verified by algorithm simulation and platform simulation. Results: Compared with RRT*, shorter path and high node utilization are shown in the improved RRT*, which cut down about 1mm in the average path length and increased about half in the average node utilization. In the meantime, the fitting results are accepted, and the results of algorithm simulation and platform simulation showed good consistency and feasibility. Conclusions: This study revealed that the improved RRT* was superior to RRT*, and the proposed method could be used for the trajectory planning of parallel orthopedic robots, which has some significance for bone fracture and deformity correction.

MATEC Web of Conferences, open access proceedings in Materials science, Engineering and Chemistry