ADVERTISEMENT

Analyzing Optimal Grasping And Passing Techniques For Robots During “Handover” Tasks

When we think about robots, we imagine humanoid machines that are capable of doing almost everything we can do, and probably more. The reality of things at the moment is slightly different.

Robots are effectively used inside factories that are built in a way to accommodate repetitive and well-calculated pre-programmed robotic motions. However, the newest wave of industrialization, the so-called Industry 4.0, envisions spaces shared by human workers and robots that efficiently collaborate to increase the productive output while also increasing the safety of the human workers.

ADVERTISEMENT

Such collaborations are meant to help human workers during physically demanding tasks. In this perspective, robots are naturally expected to interact with their human colleagues. These interactions are even more welcome in environments like houses or hospitals, where robots could help in retrieving bulky items and in caring for the patients, respectively. In all of these scenarios, robots are awaited to perform tasks involving object manipulation such as grasping a mug from a house cabinet, retrieving a specific medicine from a hospital cabinet to give it to a patient, or passing over a screwdriver to a human worker in a manufacturing factory.

A handover is a clear example of an action that is expected to happen in such shared environments. A handover is the action of an agent (passer) passing an object on to another agent (receiver). The passer usually takes hold of the object and then hands it over to the receiver. The receiver instead is interested in taking the object from the passer in order to use it for their own task. In this perspective, such collaborative effort is an intricate process of prediction and interpretation of the other agent’s actions – targeting an adjustment if the predictions are violated by unexpected actions from the other agent. Intuitively, passers need to release the objects as naturally as possible: for example, holding on to the object for too long could be confusing to the receiver as to the real intention of the passer to give the object.

On the other hand, receivers want to get the objects from the passers and get on with their tasks as quickly as possible. An example: imagine to have to hand over a pen to a colleague. Knowing that our colleague will use the pen to sign a document, we will most possibly grasp and hand over the pen from the tip. This action leaves the body of the barrel of the pen as unencumbered as possible, for the receiver to grasp the pen and be immediately able to sign off the document. Many influencing aspects such as non-verbal and verbal signaling (e.g., gaze and gestures) have been investigated in the human-robot interaction community; however, the role of grasp choice (“how to grasp an object”) and location (“where on the object to grasp it”) is still an open question.

Our study published in Science Robotics aims to answer the question of how the task of completing a handover changes “how to grasp an object” (grasp type) and “where on the object to grasp” (grasp location) with respect to direct use of the same object. Literature in human physiology and neuroscience has shown that an object can be grasped in multiple ways; however, the final choice is dependant on a number of factors that include also the task to use the object for, together with the size and dimension of the object and the habits of the person performing the grasp.

ADVERTISEMENT

In this perspective, our experiment has allowed us to compare grasp choices when directly performing a task to the choices of grasp made when passing objects to another person. Our conclusion is that the task of passing objects effectively changes the preferences of “how” and “where” to grasp objects. In particular, we speculate that these adjustments aim at accommodating to the receiver and increase the “easiness” with which the receiver will perform their part of the exchange and the subsequent task.

Our study has uncovered interesting strategies that we, humans, use every day – and without much thinking – when interacting with each other. We believe that a robot should possess these features should we expect robots to be able to interact with us in a natural manner. For this reason, we are of the opinion that this study is a step forward towards the presence of interactive collaborative robots in smart factories and also in common environments like houses and hospitals.

These findings are described in the article entitled On the choice of grasp type and location when handing over an object, recently published in the journal Science Robotics.

Comments

READ THIS NEXT

Atomic Clock In Space To Test Changing Speed Of Light

Look how a single candle can both defy and define the darkness — Anne Frank In 1919, a total eclipse […]

Tech Companies Lead Exodus Away From Fossil Fuels

As the costs associated with renewable energy continue to fall, tech companies like Amazon and Google are leading a transition […]

The Balance Of Parenting: How Best To Support Children’s Self-Regulation

Young children are faced with learning to control their behaviors and emotions, which is a skill labeled self-regulation. You can […]

Surface Layer Energy And Carbon Fluxes Are Vulnerable To The Representation Of Canopy Structural And Functional Profiles

The terrestrial carbon uptake currently accounts for about one-third of the annual global carbon sink in the atmosphere; however, future […]

Effects Of Acute Suppression Of Slow-Wave Sleep On Young Men’s Androgen Levels

Sleep restriction and sleep disturbances are increasingly common in modern industrialized societies, due to, for instance, work-related stress and atypical […]

Kingdom Plantae: Characteristics And Examples

Taxonomy in biology refers to the science of classifying and organizing living organisms into groupings based on shared characteristics. Organisms […]

What Is The Rock Cycle: Definition, Diagram, And Examples

The rock cycle is a concept of geology that describes the transition of rocks between the three rock types: igneous, […]

Science Trends is a popular source of science news and education around the world. We cover everything from solar power cell technology to climate change to cancer research. We help hundreds of thousands of people every month learn about the world we live in and the latest scientific breakthroughs. Want to know more?