Creating intelligent social robots
For all his language skills, rationality and attentiveness, Star Wars humanoid C-3PO has never won as many hearts as the zany R2-D2.
The reasons behind this may be more important than you think. A new generation of robots is coming to teach our children, treat our sick and uphold our laws. For this to work, we humans will need to trust our mechanical helpers. But how can we be sure they'll be lovable like R2-D2, and not annoying like C-3PO?
A collection of research articles by scientists from leading tech institutions sheds light on human-robot interaction (HRI): an emerging field investigating the highly nuanced relationship between people and machines. Selected as a 2018 Spotlight finalist, Affective and Social Signals for HRI is helping to inform the design of robots with the looks, social skills and personality to successfully move into our hospitals, work places and homes.
With 55,000 article views and 4,000 article downloads to date — as well as media mentions in the BBC and other major outlets — it’s clear the topic has widespread relevance and interest.
Dr Hatice Gunes explains why she and fellow computer scientists Dr Ginevra Castellano and Prof Bilge Mutlu launched this Research Topic, some of its outcomes, and what winning the 2018 Frontiers Spotlight Award would mean for this emerging and exciting field.
Why was this Research Topic created?
"Social robots have enormous potential for public good — and are poised to transform business, society and the global economy in ways that are difficult to imagine,” says Gunes.
“It’s important they have the right skills to provide both physical and social support to human users, and to sustain productive long-term relationships with us in a variety of settings.”
However, she admits the capabilities of current social robots are quite limited. There's also a major gap between public perception of robot capabilities and their actual capabilities.
“Robots are already starting to take on social roles in homes, workplaces, and public spaces — but the public, largely misguided by sci-fi movies, has skewed opinions and unrealistic expectations of humanoid robots."
Designing robots with socio-emotional skills is a challenging research area still in its infancy.
"Our aim was to bring together the latest works and developments from across a range of research groups and disciplines working in relevant fields and foster further interactions and collaborations," says Gunes. "My previous experience is that launching a research article collection on a specific theme is the best way of doing this."
What type of research is included?
Affective and Social Signals for HRI narrows the gap between sci-fi and reality in human-robot interaction.
The look, skills, personality and assessment of future social robots are imagined at a new level of detail in experiments and observations on nonverbal behaviors — cues like interpersonal distance, posture, gestures, gaze, vocal outbursts and silences — which are critical in human interactions.
“Truly intelligent social robots need to be able to interpret and express these nonverbal cues to participate in high-level phenomena like first-impression formation, social roles, interpersonal relationships, mood, emotions and personality.”
What are some outcomes of this Research Topic?
Human-robot interaction is developing at the intersection of many disciplines. Promoting and facilitating this interdisciplinary collaboration is another key contribution of the Research Topic towards the goal of socially intelligent robots.
“Our foremost achievement in launching this Frontiers Research Topic is to foster further interactions and collaborations between the fields of psychology, nonverbal behavior, computer vision, social signal processing, affective computing, and HRI.”
Gunes sees another key success in the huge audience reached by the Research Topic.
“We believe that this contributes to creating a realistic public view regarding the capabilities and limitations of social robots.
“It is also expected to help with technology adoption, for instance, the use and acceptance of social robots as new personal products for the public good.”
The research collection has generated a number of interesting findings.
“We usually imagine the ideal robot to have flawless capabilities for perception, reasoning and action.
“However, the article ‘To Err Is Robot’ by Mirning et al. showed surprising results — people liked a faulty robot significantly better than one that interacted flawlessly.
“This and other findings will help improve the capabilities of the humanoid robots — and also the acceptance of social robotics and autonomous systems as new personal products.”
Why should this Research Topic win the 2018 Spotlight Award?
The impact of these findings is already being seen, with extensive readership and several citations.
“The most immediate application of this Research Topic is in telepresence robotics, and guiding future HRI research,” explains Gunes.
But winning the 2018 Frontiers Spotlight Award — with a prize of US $100,000 to fund a conference on the topic — would enable these achievements to be pushed to new levels, she argues.
“Ultimately social robots are a disruptive technology. In the near future, they are expected to become as indispensable as mobile phones in today’s society.
“Even in the short-to-medium term, this line of research could see industry-transforming application in food, retail and entertainment, as well as healthcare and education.
“A conference on this Research Topic would provide the visibility needed to bring together the most prominent research groups and industry partners working in the relevant multidisciplinary areas.
“This visibility would also encourage further investment, recognition and dissemination of important findings, and recruitment of new talent – enabling the creation of an ecosystem of ambitious students, researchers and industrialists who will ensure the future growth of social robotics.”
About the Frontiers Spotlight Award
The annual Frontiers Spotlight Award supports emerging and important fields of research published as a Research Topic in Frontiers journals. The winning team of Topic Editors receives US$100,000 to organize an international scientific conference on the theme of their successful Research Topic. Learn more about the Award
About Frontiers Research Topics
Research Topics are peer-reviewed article collections published around specialized themes. They bring together leading researchers from different institutions, locations and fields of interest, who collaborate and contribute articles.
Published on Frontiers’ award-winning platform, Research Topics are fully open and accessible, and become highly visible collections of work, enhancing both the readership and citations of articles. Learn more about Research Topics
How to enter
Every Research Topic that closes within the award period and is completed with at least 10 articles, will be considered for the Spotlight Award. Shape your field and have the chance to organize your own conference – suggest your Research Topic today!
Research Topic Editors