Motion Primitive Discovery in a Humanoid Robot via Self-Organising Maps for Phase Recognition

2026-07-21

AI Summary

This work proposes a two-layer computational architecture inspired by the mirror neuron system for online action-phase recognition and self-organized discovery of motor primitives in humanoid robots. The first layer employs two self-organizing maps (SOMs) to learn topological representations of arm and hand kinematics, respectively. The second layer utilizes an echo state network (ESN) to model the temporal dynamics of SOM activation trajectories, enabling high-accuracy online phase identification. Experiments on the NICO simulation platform demonstrate that the SOMs effectively encode complementary motion features, and their activation trajectories preserve discriminative structural information critical for distinguishing action phases. These results validate the efficacy of the proposed hierarchical architecture in integrating motion representation with temporal modeling.

Abstract

Understanding the computational basis of action recognition is a central challenge in social cognition as well as in human-robot interaction. Inspired by the Mirror Neuron System (MNS), we propose a two-level architecture for motor primitive discovery and online phase recognition applied to the NICO humanoid robot. At the first level, two Self-Organising Maps (SOMs) learn topographic representations of arm kinematics (A-SOM) and hand kinematics (H-SOM) from simulated trials covering seven motor actions. The maps are trained on non-redundant features identified through hierarchical correlation analysis of motion trajectories. The results show that the two SOMs encode complementary aspects of motor behaviour. At the second level, an Echo State Network (ESN) evaluates whether temporal trajectories of SOM activations, represented by consecutive best-matching units, are sufficient for online recognition of the currently executed movement phase. The results show that SOM-based trajectories preserve the dominant phase-discriminative structure of the movement, while contextual information provides only a secondary refinement. Our contribution is the integration of established SOM and ESN methods within an MNS-inspired architecture for motor primitive representation and online phase recognition. The results are compatible with the computational hypothesis that self-organised motor representations, when temporally integrated, can support accurate online recognition of ongoing movement phases.

Problem

• motion primitive discovery
• phase recognition
• humanoid robot
• action recognition
• mirror neuron system

Innovation

• Self-Organising Maps
• Echo State Network
• Motor Primitive Discovery
• Phase Recognition
• Mirror Neuron System