Variable Impedance Actuators (VIA) have received increasing attention in recent years as many novel applications involving interactions with an unknown and dynamic environment including humans require actuators with dynamics that are not well-achieved by classical stiff actuators. This paper presents an overview of the different VIAs developed and proposes a classification based on the principles through which the variable stiffness and damping are achieved. The main classes are active impedance by control, inherent compliance and damping actuators, inertial actuators, and combinations of them, which are then further divided into subclasses. This classification allows for designers of new devices to orientate and take inspiration and users of VIA's to be guided in the design and implementation process for their targeted application.

How do physical interactions with others change our own motor behavior? Utilizing a novel motor learning paradigm in which the hands of two - individuals are physically connected without their conscious awareness, we investigated how the interaction forces from a partner adapt the motor behavior in physically interacting humans. We observed the motor adaptations during physical interactions to be mutually beneficial such that both the worse and better of the interacting partners improve motor performance during and after interactive practice. We show that these benefits cannot be explained by multi-sensory integration by an individual, but require physical interaction with a reactive partner. Furthermore, the benefits are determined by both the interacting partner's performance and similarity of the partner's behavior to one's own. Our results demonstrate the fundamental neural processes underlying human physical interactions and suggest advantages of interactive paradigms for sport-training and physical rehabilitation.

This paper presents a novel human-like learning controller to interact with unknown environments. Strictly derived from the minimization of instability, motion error, and effort, the controller compensates for the disturbance in the environment in interaction tasks by adapting feedforward force and impedance. In contrast with conventional learning controllers, the new controller can deal with unstable situations that are typical of tool use and gradually acquire a desired stability margin. Simulations show that this controller is a good model of human motor adaptation. Robotic implementations further demonstrate its capabilities to optimally adapt interaction with dynamic environments and humans in joint torque controlled robots and variable impedance actuators, without requiring interaction force sensing.

From a parent helping to guide their child during their first steps, to a therapist 15 supporting a patient, physical assistance enabled by haptic interaction is a fundamental modus 16 for improving motor abilities. However, what movement information is exchanged between 17 partners during haptic interaction, and how this information is used to coordinate and assist 18 others remains unclear 1 . Here, we propose a model where haptic information, provided by touch 19 and proprioception 2 , enables interacting individuals to estimate the partner's movement goal, 20 and to improve their own motor performance. We utilize an empirical physical interaction task 3 21 to show that our model can explain human behaviours better than existing models of interaction 22 in literature [4][5][6][7][8] . Furthermore, we experimentally verify our model by embodying it in a robot 23 partner and checking that it induces the same improvements in motor performance and learning 24 in a human individual as interacting with a human partner. These results promise collaborative 25 robots that provide human-like assistance, and suggest that movement goal exchange is the key 26 to physical assistance. 27 Humans are adept at physically interacting with and assisting one another, from helping 28 children to walk, to the incredible feats of balance in acrobatics, and synchrony during the 29 Tango. For over a decade, physical coupling has been documented to promote partners to adopt 30 specialised roles 9-11 and enable pairs or dyads to improve in many joint tasks 3,9,12,13 . However, 31 the underlying computational principle that enables movement coordination is still unknown 1 . 32 The improvement in interacting partners has been shown not to be due to changes in attention or 33 impedance of the interacting limbs 3 , and is absent when the interaction is not physical 14,15 and 34 when the interacting partners do not fully share control 16 . These results highlight the importance 35 of haptics, the sensory modality related to tactile and proprioceptive senses 2 , during continuous 36 physical interactions, and suggest that individuals jointly coordinate with a partner by 37 exchanging information haptically. However, what information is being exchanged and how it 38 is used to adapt one's behaviour remains unknown. We hypothesised that haptically interacting 39 partners can estimate and exchange sensory information about the task goal and the uncertainty 40 of this information with their partner, and tested this hypothesis against competing models of 41 haptic interaction during an interactive target tracking task. 42 Specifically, we first simulated the mechanical dynamics and control behaviour of 43 2 individual partners during the interactive task. We considered our proposed interpersonal goal 44 integration model against three well-known models of interaction in literature that propose 45 different information being exchanged between the partners. We compared the prediction of 46 these...

To move a hard table together, humans may coordinate by following the dominant partner’s motion [1–4], but this strategy is unsuitable for a soft mattress where the perceived forces are small. How do partners readily coordinate in such differing interaction dynamics? To address this, we investigated how pairs tracked a target using flexion-extension of their wrists, which were coupled by a hard, medium or soft virtual elastic band. Tracking performance monotonically increased with a stiffer band for the worse partner, who had higher tracking error, at the cost of the skilled partner’s muscular effort. This suggests that the worse partner followed the skilled one’s lead, but simulations show that the results are better explained by a model where partners share movement goals through the forces, whilst the coupling dynamics determine the capacity of communicable information. This model elucidates the versatile mechanism by which humans can coordinate during both hard and soft physical interactions to ensure maximum performance with minimal effort.

No abstract

Ganesh G, Haruno M, Kawato M, Burdet E. Motor memory and local minimization of error and effort, not global optimization, determine motor behavior. J Neurophysiol 104: 382-390, 2010. First published May 19, 2010 doi:10.1152/jn.01058.2009. Many real life tasks that require impedance control to minimize motion error are characterized by multiple solutions where the task can be performed either by co-contracting muscle groups, which requires a large effort, or, conversely, by relaxing muscles. However, human motor optimization studies have focused on tasks that are always satisfied by increasing impedance and that are characterized by a single erroreffort optimum. To investigate motor optimization in the presence of multiple solutions and hence optima, we introduce a novel paradigm that enables us to let subjects repetitively (but inconspicuously) use different solutions and observe how exploration of multiple solutions affect their motor behavior. The results show that the behavior is largely influenced by motor memory with subjects tending to involuntarily repeat a recent suboptimal task-satisfying solution even after sufficient experience of the optimal solution. This suggests that the CNS does not optimize co-activation tasks globally but determines the motor behavior in a tradeoff of motor memory, error, and effort minimization.

No abstract