Human manipulation of objects connects the senses of sight and touch from an early age, and this sensory connection is strengthened over the course of child development.
Manipulation of unknow objects is a topic of interest in the field of robotics We expect robots to be capable of recognizing everyday objects in the environment to provide us support in daily tasks, such as tiding up our homes, handing over working materials at our offices or helping us with our grocery shopping, to name a few examples. Hence, the ability to distinguish between a variety of different objects, including a learning process of unknown objects, is a significant case of study in this regard.
In this project, we study to what extent cross-modal perception, connecting the senses of sight and touch, can substantially benefit a robots’ cognitive system ability to seamlessly adapt to changes in the environment, increasing their autonomy and understanding of the world through multimodal learning.
We also use this expertise to raise awareness of the capabilities and limitations of robots among the general public, helping to tackle one of the biggest challenges preventing the wide-spread use of social robots: a mismatch between what the robot can actually do and what the user expects the robot should do.
Financiado el Ministerio de Ciencia, Innovación y Universidades del Gobierno de España - Agencia Estatal de Investigación con la ayuda nº PID2020-119725GA-I00 (PID2020-119725GA-I00 / AEI / 10.13039/501100011033) para el proyecto "Percepción cross-modal auto-adaptativa en robótica cognitival" (CMES)