Doktorandské kolokvium KAI - Radovan Gregor (1.12.2025)
v pondelok 1.12.2025 o 13:10 hod. v miestnosti I 9
Prednášajúci: Radovan Gregor
Názov: Mirroring mechanisms for trustworthy human-robot interaction
Termín: 1.12.2025, 13:10 hod., I/9
Abstrakt:
The discovery of a group of neurons in the premotor cortex of macaque monkeys that became active both when the monkeys performed certain actions and when they observed another individual performing the same actions proposed the existence of a neural system that "mirrors" the actions of others,giving rise to a name Mirror Neuron System (Rizzolatti G., 1992). This discovery has had a major impact on our understanding of the neural basis of social cognition and has been the subject of numerous studies.
In this talk we will first focus on mirroring-based prediction of motor intentions using an Echo State Network in a simulated robotic environment. Predicting motor intentions is crucial for understanding the observed behavior. We approach this task aiming at predicting future states, and ultimately, the target object, of a moving robotic arm based on mirroring. Then we will talk about a Self-Organizing Map model trained on sliding windows of joint positions, aperture, and palm orientation from our robot hand logs, using grasp labels (Power / Precision / Side). The SOM learned a 2D topographic map where different regions specialize for different grasp types. At test time, we classify an unknown trajectory segment by finding its best-matching unit on the map and reading out the dominant grasp type from the learned label distribution.