Object Handovers: a Review for Robotics

Valerio Ortenzi, Akansel Cosgun, Tommaso Pardi, Wesley Chan, Elizabeth Croft, Dana Kulić

Seminal July 2020

TL;DR A survey of human-robot object handovers that splits each exchange into a pre-handover phase and the physical handover, and compares 51 papers with real-robot experiments. It proposes a minimal evaluation set: success rate, total handover time and the receiver's task time, plus Likert questions on fluency, trust and working alliance.

What they do

A handover is a joint action in which a giver passes an object to a receiver. This review splits it into a pre-handover phase (communication, grasping and transport, until the receiver first touches the object) and the physical handover (until the giver lets go). It covers robot-to-human (R2H) and human-to-robot (H2R) direct handovers and draws on human-human studies.

How it works

Pre-handover. Speech, gaze and body-pose cues; grasps that leave the receiver the part they need next (a screwdriver offered by its handle); predicting the partner’s motion; choosing the handover location; legible, reactive motion.
Physical handover. Grip-force modulation and release timing: in human studies the giver keeps the object safe and the receiver sets the timing. Also error handling, such as telling a pull from an accidental bump.
Safety and metrics. Safety of human, object and robot; task performance, psycho-physiological and questionnaire metrics.
Tables. 51 real-robot papers compared by paradigm, phase, sensors, handover location, post-handover task, metrics and number of test objects.

Results

R2H is studied more than H2R, and the pre-handover phase more than the physical exchange. Handover locations are mostly fixed or pre-planned, and few papers include communication.
Most robot givers release the object fully once they detect a pull.
Most studies use one object class, usually bottles or boxes; the largest sets are 22 and 26 objects. Very few give the receiver a task after the handover.
Proposed protocol: success rate, total handover time and receiver task time; eight Likert items on fluency, trust and working alliance; a post-handover task; objects eliciting power, intermediate and precision grasps.

Limitations

A review with no new experiments, limited to direct handovers. The proposed metrics leave out psycho-physiological measures, which need body-worn sensors, and the protocol is not tested.

Figures

From Figure 1: A direct handover in the physical exchange phase. A robot passes a mustard bottle to a human, and both... — From Figure 1: A direct handover in the physical exchange phase. A robot passes a mustard bottle to a human, and both hands are on the object.

From Figure 2: The giver's tasks (object stability, presentation, release) and the receiver's tasks (acquisition, sta... — From Figure 2: The giver's tasks (object stability, presentation, release) and the receiver's tasks (acquisition, stability, use) across the phases of a handover, from initiation to the receiver using the object.

From Figure 3: Handover metrics, grouped into task performance (timing, success rate, interaction force) and user exp... — From Figure 3: Handover metrics, grouped into task performance (timing, success rate, interaction force) and user experience (psycho-physiological and subjective measures).

From Figure 4: Real-robot handovers from earlier work: two robot-to-human setups (left, middle) and one human-to-robo... — From Figure 4: Real-robot handovers from earlier work: two robot-to-human setups (left, middle) and one human-to-robot setup (right).

Abstract

This article surveys the literature on human-robot object handovers. A handover is a collaborative joint action where an agent, the giver, gives an object to another agent, the receiver. The physical exchange starts when the receiver first contacts the object held by the giver and ends when the giver fully releases the object to the receiver. However, important cognitive and physical processes begin before the physical exchange, including initiating implicit agreement with respect to the location and timing of the exchange. From this perspective, we structure our review into the two main phases delimited by the aforementioned events: 1) a pre-handover phase, and 2) the physical exchange. We focus our analysis on the two actors (giver and receiver) and report the state of the art of robotic givers (robot-to-human handovers) and the robotic receivers (human-to-robot handovers). We report a comprehensive list of qualitative and quantitative metrics commonly used to assess the interaction. While focusing our review on the cognitive level (e.g., prediction, perception, motion planning, learning) and the physical level (e.g., motion, grasping, grip release) of the handover, we briefly discuss also the concepts of safety, social context, and ergonomics. We compare the behaviours displayed during human-to-human handovers to the state of the art of robotic assistants, and identify the major areas of improvement for robotic assistants to reach performance comparable to human interactions. Finally, we propose a minimal set of metrics that should be used in order to enable a fair comparison among the approaches.

Summaries are written automatically from each paper's text. Figures and videos come from the paper and its project page; all credit goes to the authors.

Matt Strong