Stefan Escaida ● Profesor Asistente

Fondos para apoyar la realizacion de la Third Latin American Summer School on Cognitive Robotics (LACORO 2024). La primera edición se realizó online en octubre de 2020; la segunda fue presencial en enero de 2023; la tercera versión tendrá lugar entre el 9 y el 13 de diciembre de 2024 en la Universidad de O'Higgins en Rancagua, Chile. https://lacoro.org/ Esta Escuela de Verano beneficiará principalmente a Estudiantes y Académicos de las Américas interesados en la Investigación en Inteligencia Artificial aplicada a la Robótica. Nuestro objetivo es fomentar la colaboración nacional y regional en esta área de investigación. Para la primera edición, alcanzamos 241 inscripciones para actividades online de todo el mundo, y la segunda versión tuvo 166 inscripciones para actividades presenciales en enero de 2023, principalmente de Chile, México, Argentina, Brasil y Uruguay.

In the previous project the research focus was on the sensing principle and the development of prototype modules for a tactile proximity sensor (TPS). In the current project the focus is on the methods and algorithms with which the events in the near proximity of the robots can be modelled by means of these sensors. Collectively, the TPS on the robot surface and gripper constitute a smart skin. The application scenarios here are the Active Exploration of the Environment, Grasping and the Safe Human-Robot-Interaction. The methods to be developed will improve significantly on the quality of state of the art methods and expand the horizon of possible solutions for these problems. The capacitive measuring principle and the spatial resolution in both, the tactile and proximity modalities, enable an area-wide and distance based coverage of the robot surroundings. It is the first time that algorithms for Exploration, Grasping and Safe-Human-Robot-Interaction are presented that rely simultaneously on both tactile and proximity sensing with spatial resolution. The goal of the Exploration is to research which methods and strategies enable the robot to acquire a contact- and proximity-based world-model by means of TPS. The quality of state of the art solutions for grasping should be improved significantly. For the Safe-Human-Robot-Interaction algorithms should be developed that adapt the robots path and velocity according to the current situation as determined by the TPS. Also, new algorithms should make a TPS-based Interaction possible with the goal of robot controlling and programming by the human through tactile and proximity input. Finally, according to the context of the task and situation at hand (Exploration, Manipulation and Interaction) the robot should show an appropriate behaviour which is given by a behavioural strategy that will also be developed. The starting point in the project is the TPS-modules which were successfully developed in the previous project. At the beginning, research will be focused on the algorithms for signal processing that extract robust features from TPS for higher level tasks. This step is followed by the integration of TPS into a robot system realizing the smart skin. Building on these steps the methods for Exploration, Manipulation and Safe-Human-Interaction are developed. Finally, the results of the project will be shown and evaluated in a combined demonstration scenario that includes a robot endowed with TPS.

Desarrollo de réplicas fidedignas e interactives de objetos deformables del mundo real. La principal aplicación estudiada serán los phantomas de órganos humanos, que permitirán la validación de intervenciones así como ayudar durante la formación de especialistas.