Integrating Perception and Emotion for Social Robotics

Social Robotics Driven by Intelligent Perception and Endogenous Emotion-Motivation Core

  • Wanyue Jiang, Qingdao University, Qingdao City, China (email: jwy@qdu.edu.cn)
  • Xiaorui Liu, Qingdao University, Qingdao City, China

The harmonious and natural social interaction with people is the most pursued of robot intelligence research, it also has been regarded as a sign that human society has entered the era of human-machine fusion. At present, the research on the theory and technology of robot social interaction has made some progress in short-term and specific scenarios, but a social interaction theory system that integrates social knowledge, intention understanding, social motivation generation, and behavioral expression control has not yet been formed. In this session, several topics focusing on smart sensing and perception will be discussed, aiming at human-robot interaction. The session starts with the issue of robot social interaction, followed by several topics from different perspectives:

  • By exploiting how to build the social protocol and knowledge
  • By realizing the social perception based multi-modalities
  • By modeling robot emotional core and motivation
  • By designing the expression of robot social behavior

Adaptive Human-Robot Interaction with Brain-Inspired Models

AdaPtive beHavioRal mODels of robotic systems based on brain-inspired AI cogniTivE architectures (APHRODITE)

  • Laura Fiorini, Università degli studi di Firenze, Italy (email: laura.fiorini@unifi.it)
  • Anna Esposito, Università della Campania “Luigi Vanvitelli”, Italy
  • Praminda Caleb-Solly, University of Nottingham, United Kingdom

The cooperation between humans and robots is becoming increasingly important in our society. Consequently, there is a growing interest in the development of models that can enhance the interaction between humans and robots. A key challenge in the Human-Robot Interaction (HRI) field is to provide robots with cognitive and affective capabilities, developing architectures that let them establish empathetic relationships with users. The motivation behind developing a robot with social behavior is to enable it to have the social intelligence and awareness for situationally appropriate reactions, responses, and functionality, thus engendering trust and smooth human-robot interactions.  Up to now the development of robotic functionality has mainly been devoted to transmitting the content of information without paying much attention to “how” this is tailored to specific human preferences and behaviors. The robot should be able to process all the input data it gathers from the sensors but also combine it intelligently with the information provided by clinicians and other professionals to accurately profile the user and provide the appropriate behaviors. From the application perspective, the social robot should be usable in rehabilitative/assistive contexts to propose personalized care sessions and provide feedback to clinicians. In summarizing, the main pillars of APHRODITE are:

  • USER PERCEPTION and MODELLING: By incorporating a multimodal perception system and feedback from other sources (e.g. clinicians) the robot is able to build and maintain an up-to-date profile of the user and to propose the appropriate behavior for interaction, e.g. as part of personalized assistance or rehabilitation plans. 
  • INTERACTION: using a dynamic user profile, the robot should be able to tailor personalized interaction by relying on typical human-human communication channels to foster empathy and engagement.
  • THERAPY/REHABILITATION: In addition to a personalized interaction, the robot should be able to propose, among other activities, rehabilitative intervention therapy such as cognitively/physically stimulating and supporting humans in doing some activities acting as a companion/coworker. 

While these research topics are potentially relevant with a high social and scientific impact, there are still gaps and challenges from a scientific perspective. This requires bringing the research community together to prompt more discussion and development, and sharing experiences of how to carry out extensive testing in the field for improving robot capabilities and clinically validating solutions for healthcare applications. In this context, this workshop aims to: present the latest research on the development of behavioral models for robot solutions; investigate how social robots can be used to support assessment, rehabilitation, and therapy practices; investigate how they can be informed by and contribute to our understanding of relevant theories of (social) cognition; investigate the social cues to be included in these systems; investigate how to tailor and personalize their HRI capabilities; investigate the role of the clinicians and the therapists in user profiling. Results from preliminary field tests will also be reported.