Description
SOMA (Socio-Physical Model of Activities) is an OWL-based robotics domain ontology for modeling everyday human activities and their associated objects, contexts, and roles. Developed as part of the EASE research initiative, SOMA provides a structured vocabulary and formal axioms that describe how objects are involved in actions, how activities are composed, and how environments constrain or afford behavior.
SOMA is designed to serve as a shared semantic framework for autonomous systems and cognitive architectures. It supports symbolic representation of activities, object use, spatial configurations, and event structure — all grounded in a way that is accessible to reasoning systems, including those used in robotics.
While SOMA itself is a symbolic, logic-based ontology, it is often used in conjunction with reasoning systems (e.g., KnowRob) that can infer higher-level interpretations from observed data. These systems use SOMA as a semantic substrate, but the reasoning takes place in those external frameworks.
Why SOMA
Autonomous agents operating in human environments need rich symbolic knowledge about objects and activities — not just their geometry, but how they are typically used, in what roles, and within what context.
SOMA provides:
- Conceptual building blocks for modeling composite activities and their participants
- Role-based modeling of objects and agents in task contexts (e.g., knife as a tool)
- Tempo-spatial structure of events and situations (e.g., preconditions, sequences)
- Linkability with perception and reasoning systems through formal semantics
Key Features
- OWL 2 ontology with clear modular structure (e.g., for activities, objects, roles, contexts)
- Designed to be extendable and interoperable with domain-specific vocabularies
- Encodes common-sense knowledge useful for interpreting everyday manipulation tasks
Example Video
Software Components
SOMA is composed of ontology files that can be downloaded from GitHub.
The page also provides versioned releases of SOMA.
Courses and Tutorials
Modelling with SOMA. Robin Nolte, Robert Porzel. EASE Fall School 2023 Tutorial PDF
Authors and Contact Details
• Dr. Daniel Beßler
Tel: +49 421 218 64016
Email: danielb@cs.uni-bremen.de
Profile: https://ai.uni-bremen.de/team/daniel_beßler
• Dr. Robert Porzel
Tel: +49 421 218 64407
Email: porzel@tzi.de
Profile: https://www.uni-bremen.de/dmlab/team/dr-ing-robert-porzel
• Dr. Mihai Pomarlan
Email: mihai.pomarlan@uni-bremen.de
• Sascha Jongebloed
Tel: +49 421 218 64008
Email: jongebloed@uni-bremen.de
Profile: https://ai.uni-bremen.de/team/sascha_jongebloed
• Robin Nolte
Tel: +49 (0)421 218-64416
Email: nolte@uni-bremen.de
Profile: https://www.uni-bremen.de/dmlab/team/robin-nolte
Publications
Daniel Beßler, Robert Porzel, Mihai Pomarlan, Abhijit Vyas, Sebastian Höffner, Michael Beetz, Rainer Malaka and John Bateman, “Foundations of the Socio-physical Model of Activities (SOMA) for Autonomous Robotic Agents”, In Formal Ontology in Information Systems - Proceedings of the 12th International Conference, FOIS 2021, Bozen-Bolzano, Italy, September 13-16, 2021, IOS Press, 2021, doi: 10.3233/FAIA210379