On the Role of Physical Interaction on Performance of Object Manipulation by Dyads

Abstract

Human physical interactions can be intrapersonal, e.g., manipulating an object bimanually, or interpersonal, e.g., transporting an object with another person. In both cases, one or two agents are required to coordinate their limbs to attain the task goal. We investigated the physical coordination of two hands during an object-balancing task performed either bimanually by one agent or jointly by two agents. The task consisted of a series of static (holding) and dynamic (moving) phases, initiated by auditory cues. We found that task performance of dyads was not affected by different pairings of dominant and non-dominant hands. However, the spatial configuration of the two agents (side-by-side vs. face-to-face) appears to play an important role, such that dyads performed better side-by-side than face-to-face. Furthermore, we demonstrated that only individuals with worse solo performance can benefit from interpersonal coordination through physical couplings, whereas the better individuals do not. The present work extends ongoing investigations on human-human physical interactions by providing new insights about factors that influence dyadic performance. Our findings could potentially impact several areas, including robotic-assisted therapies, sensorimotor learning and human performance augmentation.

Let's Move It Together: A Review of Group Benefits in Joint Object Control

“…In sum, the majority of studies investigating joint tasks with distributed object control find that individuals outperform dyads (Knoblich and Jordan, 2003 ; Bosga and Meulenbroek, 2007 ; Newman-Norlund et al, 2008 ; Mojtahedi et al, 2017 ). Findings also indicate that joint performance depends on (1) whether co-actors receive specific information about each other's actions (beyond seeing their combined effects on the controlled object) (Knoblich and Jordan, 2003 ; van der Wel et al, 2011 ); (2) the degree of coordination required (e.g., coordination in one or two spatial dimensions); (3) the type of control required (i.e., a coarse or fine type of control) (Wahn et al, 2016 ); and (4) co-actors' interindividual skill differences (Wahn et al, 2016 ; Mojtahedi et al, 2017 ).…”

“…The similarity in co-actors' individual performance levels has been shown to predict group benefits in the two-dimensional object control task described above (Wahn et al, 2016 ): The more similar the co-actors' individual skills, the higher the group benefit when they perform together. There is also evidence that individuals do not benefit equally from interpersonal coordination (Mojtahedi et al, 2017 ). In particular, when two co-actors physically lifted and balanced an object by each grasping one of the two handles of the object, only the “worse” co-actor benefited (relative to her individual bimanual performance) whereas the “better” co-actor's performance tended to decrease when performing the task jointly (Mojtahedi et al, 2017 ).…”

“…There is also evidence that individuals do not benefit equally from interpersonal coordination (Mojtahedi et al, 2017 ). In particular, when two co-actors physically lifted and balanced an object by each grasping one of the two handles of the object, only the “worse” co-actor benefited (relative to her individual bimanual performance) whereas the “better” co-actor's performance tended to decrease when performing the task jointly (Mojtahedi et al, 2017 ). However, in line with Bosga and Meulenbroek ( 2007 ), the joint performance was still worse than the individual performance in this type of control task (Mojtahedi et al, 2017 ).…”

“…So far, a number of factors that have been investigated in distributed control have not yet been investigated in redundant control. In particular, it remains to be tested whether co-actors' performance in redundant control tasks is affected by the degree of coordination (e.g., coordination in one or two spatial dimensions), by the type of control required (i.e., a coarse or fine type of control) (Wahn et al, 2016 ), and by co-actors' interindividual skill differences (Wahn et al, 2016 ; Mojtahedi et al, 2017 ). Future research could also investigate how much time co-actors typically need to voluntarily distribute control, and whether the type of control distribution varies across dyads and across time.…”

The Need for Combining Implicit and Explicit Communication in Cooperative Robotic Systems

“…Some of the processes listed above have suggested how these coordination features can be used in human-robot and robot-robot interactions, particularly in the field of human rehabilitation (Ganesh et al, 2014 ; Mojtahedi et al, 2017a , b ; Sawers et al, 2017 ). Other processes have already been investigated within the context of human-robot interaction.…”