We propose, implement, and evaluate a natural human-machine control interface for a variable stiffness transradial hand prosthesis that achieves tele-impedance control through surface electromyography (sEMG) signals. The famous trade-off between stability and transparency while the robot is coupled with an environment was then study thanks to the obtained estimations, to enhance the tuning of the endpoint admittance control empirically designed. The estimations of the arm's virtual trajectories both in position and force relied on spline interpolations and sine optimisations, for small deviations that did not alter the performances of the task.A cohort of participants took part in experiments proposed to observe significant variations of the viscoelastic apparent parameters, and improve the understanding of the implications of such variations during a physical interaction with a robot.
A methodology to identify the linear parameters of the chosen impedance model was designed, tackling the issue of the estimation of virtual trajectories of the arm during dynamic movements. A benchmark task was used so that rhythmic movements emerged, while haptic feedback were introduced by the robot. Perturbation rejection behaviours occuring remarkably during physical interactions can be met using this modelling.In order to estimate those properties for the human arm, an experimental test-bed was designed using an endpoint admittance controled polyarticulated robot. In this work, the focus is placed on a simplistic linear mechanical model of the human arm, with endpoint apparent parameters like stiffness, damping and mass.
Knowledge on the human endpoint vis-coelastic properties during physical interactions provides insights for the field of human movement science and also for the design of innovative bio-inspired collaborative robotic control strategies. Robots are more inclined to interact with humans or their environment for collaborative purposes.
0 kommentar(er)
