N1-[SocialSpace] - Overview

Social Learning for Cognitive Robots

Motivated by the growing number of robots deployed in human environments, the objective of project N1-[SocialSpace] is to give cognitive systems the ability to sense and act in a socially acceptable way. We develop theory and algorithms for smooth, socially compliant, safe yet task-efficient operation of cognitive systems for human-robot coexistence, interaction, and collaboration. To this end, we investigate models that describe people and groups of people with attributes, social relations, activities, and behaviors. To acquire such models we take machine learning approaches and model-based approaches that rely on findings from cognitive or social science. The models are then used as background knowledge to inform and improve key components for robots in human environments such as detection and tracking of people and groups, socially-aware task and motion planning and human-robot interaction.

More details: http://srl.informatik.uni-freiburg.de/