TRAIL at HRI2024

22. März 2024, von Tamlin Love, Ferran Gebelli, Theresa Pekarek Rosin

We are pleased to announce that two of our doctoral candidates (Ferran Gebellí and Tamlin Love) presented work at the recent ACM/IEEE International Conference on Human-Robot Interaction which took place in Boulder, Colorado, USA from 11 to 15 March.

Tamlin Love, supervised by Guillem Alenyà Ribas at the Institut de Robòtica i Informàtica Industrial, presented a late breaking report at the conference, titled Towards Explainable Proactive Robot Interactions for Groups of People in Unstructured Environments (full text). The paper presents a two-layer system for perception and decision-making for robot interaction elicitation in unstructured, dynamic and multi-person environments, and details a new method for generating explanations of decisions made by the robot, based on a counterfactual search of a causal model of the environment. A proof of concept is presented to illustrate how explanations can be found in response to different and dynamic human-robot interaction settings, and to different queries posed by a user.

Ferran Gebellí, supervised by Raquel Ros at PAL Robotics, participated in the Explainability for Human-Robot Collaboration workshop, where they presented their short paper entitled Participatory Design for Explainable Robots. The work presents a participatory design methodology to address explainability in HRI applications. The suggested method incorporates the explainability component into the co-design process, assisting in the identification of non-understandable circumstances and the testing of possible solutions at the core of the design process. This stands in contrast to merely adding explainability components into pre-existing systems. An example of the methodology's application was carried out in an intermediate healthcare facility, where we co-designed a robotic system to assist the staff in daily care routines.

Congratulations to both doctoral candidates on their achievements!