Robot-assisted acquisition of a motor skill: Evolution of performance and effort

Haptic Training: Which Types Facilitate (re)Learning of Which Motor Task and for Whom? Answers by a Review

“…-Proportional/Derivative controllers that enforce the desired position/velocity of the limb [28]- [31], -Artificial potential fields that restrict the movement to an area around the desired trajectory by correcting the movement with repulsive forces [1,2], and/or to enforce a position-dependent velocity profile [3], -Haptic guidance in force to enforce pre-recorded force profiles [32], [33], -Interactive controllers that simulate the dynamics of a (human) partner on a robot, to which the user is haptically connected during task execution [34], -Entrainment to match the frequency of a limb's motion with that of a robotic device [35], -A combination of the above -e.g., path control and derivative control [36]- [38]. As opposed to haptic guidance, in which the haptic forces are applied to reduce the movement-related errors, haptic disturbance methods apply disturbing forces to exploit the assumption that errors drive motor learning [10], [39].…”

“…When the learners' initial skill level is not taken into consideration, performance-enhancing haptic methods do not seem to provide any additional benefit over practicing the task without haptics for the learning of spatiotemporal aspects of motor tasks [28], [38], [50], [53], [75], [76]. For example, in a series of experiments that investigated the effect of different types of haptic guidance methods on learning a virtual hitting task with a joystick, the performance-enhancing haptic methods did not enhance the learning of spatiotemporal aspects [53], [75], [76].…”

“…However, to date, adaptive methods have not been extensively investigated in motor learning studies [40], [49]. Among the few studies that assessed adaptive haptic training, adaptive designs have been mostly applied on performance-enhancing methods (e.g., [38], [43], [44], [58], [75]), and rarely on performance-degrading haptic methods (e.g., [40]).…”

“…The authors attributed the first fast improvement phase to the familiarization to use the new robotic device and virtual reality visualization, and the slow phase to motor recovery, highlighting the importance of distinguishing the resultant motor (re-)learning from the familiarization phase. However, the inclusion of a familiarization phase prior to baseline or training session has been only seldom reported in the reviewed studies [26], [33], [35], [38], [41], [46], [58], [63], [64], [70], [76].…”

Robot assistance of motor learning: A neuro-cognitive perspective

“…Tamagnone et al (2012) compared the effects of haptic guidance on spatial and dynamic movement characteristics with a kind of putting task in which a virtual ball had to be putt into a virtual hole. The direction of the virtual ball depended on where the ball was hit, and the amplitude of its motion depended on the pad's velocity at the moment of impact.…”