Human-Robot Interaction

During my post-doc at GAIPS / INESC-ID I worked in the field of human-robot interaction (HRI) in the context of different collaboration projects. Overall, I developed AI and machine-learning techniques to aid the decision-making of autonomous interactive robots. In particular, in the EMOTE project I developed AI and machine learning (ML) techniques to manage the game-play and autonomous interactive behavior of a robotic tutor. In project INSIDE I created methodologies to collect data and develop the robot’s interaction behavior, in the context of HRI for children with autism disorder. In project CoWriter I prepared and help conducting Wizard-of-Oz studies with children on the impact of a robot managing a collaborative writing activity.

[bibshow file=my-publications.bib show_links=1 format=custom-ieee template=custom-bibshow highlight=”P. Sequeira”]

The works in [bibcite key=sequeira2016hri,sequeira2015cig,alves-oliveira2015aihri,alves-oliveira2016roman,ribeiro2015hri] were performed in the context of the EU FP7 EMOTE project aimed at developing novel artificial embodied tutors capable of engaging in empathic interactions with students in a shared physical space.

In [bibcite key=sequeira2016hri] a methodology is proposed for the creation of social interaction strategies for HRI based on restricted-perception Wizard-of-Oz studies (WoZ). This technique involves restricting the wizard’s perceptions over the environment and the behaviors it controls according to the robot’s inherent perceptual and acting limitations. The robot’s design lifecycle is divided into three consecutive phases: data collection, strategy extraction and strategy refinement. I created a hybrid strategy controller for the robotic tutor that includes a rule-based and a ML module. The former module encodes task information, well-known strategies and behavior patterns observed from preliminary interaction studies. The ML module learns interaction strategies from behavior demonstrated by the expert during the restricted-WoZ based on an associative metric within frequent-pattern mining. The results of an evaluation study showed that by performing restricted-WoZ studies the robotic tutor was able to engage in very natural and socially-aware interactions with real classroom students.

The work in [bibcite key=sequeira2015cig] details the implementation of a collaborative AI module for multi-player EnerCities, a turn-based environmental-awareness game where players (two young students and a robotic tutor in our scenario) with different capabilities collaboratively develop a city. The AI module that I created is capable of informing the game-playing and pedagogical decision-making of the robotic tutor by performing forward-planning to predict preferable, more sustainable, states of the city. The module also incorporates a social component that continuously models the game preferences of each student and automatically adjusts the robot’s strategy so to follow the group’s game action tendency.

The CMU-Portugal project INSIDE is aimed at developing symbiotic HRI in the context of a physical game involving children with autism spectrum disorder (ASD). The paper in describes the INSIDE system, a networked robot system designed to allow the use of mobile robots as active players in the therapy of children with ASD. The system is the first in which a fully autonomous mobile robot actively engages children with ASD during therapy in a semi-unstructured interaction. The article describes the hardware and software infrastructure that supports such rich form of interaction, as well as the design methodology that guided the development of the INSIDE system.

I also participated in the IST-EPFL collaborative project CoWriter in the context of using robots to aid in the development of children’s writing skill. The works in [bibcite key=chandra2015roman,chandra2016roman] detail a study in which groups of two children performed a collaborative learning activity involving writing words or letters on an iPad, and a robot or a human played the role of a facilitator. The results suggest that children felt more responsible over their peer’s performance in the presence of a robot facilitator. In addition, when one child acts as the teacher and the other as the learner, more corrective feedback is provided to the peer and more self-disclosure to the robot occurs.

Related Publications