The novel framework leverages Preference-Based Optimization (PBO) and considers three renowned interaction parameters: the robot's velocity profile, the separation distance between the human and the robot, and the vertical proximity to the user's head. By continuously refining these parameters based on qualitative feedback from human collaborators, the system adjusts the robot's trajectories dynamically. This feedback was used as ground truth to train machine learning models capable of predicting trust levels from human- and robot-related behavioral indicators. The overall goal of this personalization is to promote a safe, ergonomic, and trustworthy human-robot interaction environment. A chemical handling scenario was developed as a testbed for evaluating the framework.

​

The experiments were conducted under the supervision of Dr. Arash Ajoudani at the Human-Robot Interfaces and Interaction Laboratory, Italian Institute of Technology (IIT).

​

Publications:

1) Promoting trust in industrial human-robot collaboration through preference-based optimization. Outlet: IEEE RA-L (link)

2) Estimating Trust in Human-Robot Collaboration Through Behavioral Indicators and Explainability. Outlet: IEEE RA-L (link)