Source-linked AI summary
SOSA: A Lightweight Ontology for Sensors, Observations, Samples, and Actuators
Krzysztof Janowicz, Armin Haller, Simon J D Cox, Danh Le Phuoc, Maxime Lefrancois
TL;DR
Web-published sensor data increasingly needs contextual and interoperable descriptions, including information about features, properties, sampling, and actuation. The paper presents SOSA as a lightweight common core within the revised SSN, with modular extensions and alignments for broader or more rigorous uses. SOSA is designed for broad adoption while supporting representation of sensing, sampling, and actuation activities.
Problem
Reusing, integrating, searching, and interpreting Web sensor data requires context beyond raw observation results, while actuators and actuation also need a common vocabulary.
Method
The paper presents SOSA as a lightweight, extensible ontology core and explains its classes, properties, modular integration with SSN, and alignments with other specifications.
Results
SOSA provides a common core for sensing, sampling, and actuation that can serve broad users and support more rigorous axiomatization through combination with other ontologies.
Takeaways & Limitations
SOSA offers a minimal interoperability fallback for exchanging data across uses of SSN, its modules, and SOSA.
Takeaways & Limitations
SOSA does not provide classes for information objects, datasets, or catalogs, leaving cataloging and serving to other vocabularies and tools.
Abstract
from arXiv · showhide
The Sensor, Observation, Sample, and Actuator (SOSA) ontology provides a formal but lightweight general-purpose specification for modeling the interaction between the entities involved in the acts of observation, actuation, and sampling. SOSA is the result of rethinking the W3C-XG Semantic Sensor Network (SSN) ontology based on changes in scope and target audience, technical developments, and lessons learned over the past years. SOSA also acts as a replacement of SSN's Stimulus Sensor Observation (SSO) core. It has been developed by the first joint working group of the Open Geospatial Consortium (OGC) and the World Wide Web Consortium (W3C) on \emph{Spatial Data on the Web}. In this work, we motivate the need for SOSA, provide an overview of the main classes and properties, and briefly discuss its integration with the new release of the SSN ontology as well as various other alignments to specifications such as OGC's Observations and Measurements (O\&M), Dolce-Ultralite (DUL), and other prominent ontologies. We will also touch upon common modeling problems and application areas related to publishing and searching observation, sampling, and actuation data on the Web. The SOSA ontology and standard can be accessed at \url{https://www.w3.org/TR/vocab-ssn/}.
1. Introduction and Motivation
SOSA was developed as a lightweight, extensible common core for representing sensing, sampling, and actuation on the Web. It responds to growing demand for reusable, interpretable sensor data and supports both broad accessibility and more rigorous ontology alignments.
- Web reuse and integration of sensor data require context about features of interest, observed properties, sampling strategies, and related information.
- SOSA includes actuators and actuation alongside sensors, observations, procedures, and features of interest as first-class Web concepts.
- The revised SSN ontology uses vertical modules for deeper axiomatization and horizontal modules for broader scope without enriching existing term semantics.
- Modularization addresses concerns that the original SSN was too heavyweight and imposed overly strong commitments, while retaining optional DUL alignment and adding Prov-O, O&M, and OBOE alignments.
- SOSA provides a flexible framework for representing entities, relations, and activities involved in sensing, sampling, and actuation.
- SOSA is intended as a lightweight, easy-to-use vocabulary for a broad audience and as a minimal interoperability fallback across SSN uses and modules.
2. SOSA in a Nutshell
SOSA uses an event-centric, common core to represent observation, sampling, and actuation, balancing lightweight usability with extensibility and alignment to established models. Its reusable procedures, entities, and relationships support reproducible and interoperable descriptions across these activities.
- Core structure: SOSA’s event-centric perspective treats observations as acts or events, enabling a shared structure for observation, sampling, and actuation and facilitating PROV-O alignment.This differs from the original SSN pattern, where an observation represented an observation context.
- Core structure: SOSA models observations, sampling, and actuations as activities that target features, follow procedures, and are performed by objects or agents.The common activity pattern aligns with OGC Observations and Measurements and PROV-O.
- Design goals: SOSA balances expressivity, ease of use, and broad applicability, with an ALI(D) description-logic fragment efficiently supported by modern triple stores.The new SSN module yields ALRIN(D), while the older SSN used a more expressive fragment.
- Design goals: SOSA omits class-disjointness axioms so sensors and samples can themselves be features of interest, while using Schema.org annotations for informal domain and range semantics.This permits observations using sensors as well as observations about sensors.
- Procedures: Reusable procedures specify workflows, protocols, plans, algorithms, or computational methods for observation, sampling, and actuation, supporting reproducibility and semantic interoperability.For wind-speed measurement, procedure details such as standard sensor height and obstruction avoidance make results reproducible and representative.
- Observations: An observation applies a procedure to estimate an observable property of a feature or sample, involves a sensor, and yields a result represented simply or as a first-class object.First-class Result objects can carry information such as measurement units.
3. Vertical Segmentation Alignments
SOSA remains a lightweight core that can be extended through vertically segmented modules and aligned with established observation ontologies. These modules support migration, interlinking, and richer axiomatization beyond SOSA’s deliberately limited core.
- Vertical segmentation: SOSA is the core for vertically segmented modules that add higher ontological commitment through new axioms.The modules can support migration from old SSN and mapping data expressed with O&M and OBOE.
- Alignments: sosa:Observation is equivalent to o&m:Observation and is a superclass of oboe:Measurement.It should not be confused with oboe:Observation, which represents a collection of measurements of different properties of one feature of interest.
4. Using SOSA on the Web
SOSA is designed as a lightweight vocabulary for Web developers and can be embedded in common Web formats. Its examples combine SOSA with JSON-LD, RDFa, Schema.org, and PROV-O to represent observations and their context.
- Using SOSA on the Web: SOSA targets a broad audience, including Web developers accustomed to JSON serializations and Schema.org-style annotations.Schema.org supports integration across applications and data formats through HTML markup and JSON-LD.
- RDFa: RDFa 1.1 embeds SOSA observation descriptions in HTML5 and XHTML5 using attribute-level extensions.The example models another observation made by the same phone one minute later.
- SOSA and Schema.org: A Schema.org Action type adds location and participant details to a SOSA observation.The example uses Schema.org location for Canberra and can express additional action participants through schema:agent.
- JSON-LD and PROV-O: JSON-LD serializes a SOSA sensor and observation while PROV-O records the agent and role involved in sensing.The example links an observation to a sensor, observed property, result, and qualified association.
5. Modelling with SOSA
SOSA modeling guidance addresses practical grouping, cataloging, vocabulary, and interpretation choices. It also emphasizes an event-centric representation of observations, distinguishing SOSA from the original SSO-based SSN model.
- Collections: Observation collections can reduce redundancy when many observations share sensors, procedures, features of interest, or observed properties.Instead of repeating links, observations can be typed with a class such as S1Obs; SSN deployment classes can group observations by campaign.
- Cataloging and serving: SOSA does not provide information-object, dataset, or catalog classes, so DCAT and existing SSN-XG methods address cataloging, storage, serving, and streaming.This separates the lightweight SOSA vocabulary from data-management functions supplied by other recommendations and tools.
- Controlled vocabularies: SOSA leaves the choice between individuals and classes for controlled vocabularies to strategies such as SKOS, OWL2 punning, or casting patterns.The issue applies to observed and actuatable properties as well as procedures.
- Events versus records: SOSA conceives observations as acts or events rather than records or descriptions of an observation context.This supports a common structure for observation, actuation, and sampling and facilitates alignment with PROV-O, while mixed old and new SSN graphs require care.
6. Evidence of use of SOSA
SOSA and SSN were shaped by 51 use cases and 62 requirements, preserving useful SSN elements while expanding the model. Implementation evidence includes reuse across ontologies, datasets, and multiple application domains.
- Requirements: 51 use cases generated 62 requirements that informed SOSA and SSN modeling.SOSA contains 13 concepts and 21 properties, reusing concepts and properties from the original SSN alongside requested additions.
- Reuse and expansion: Three SOSA concepts and eleven properties are equivalent to corresponding terms in the original SSN.New or changed concepts cover procedures, results, observable and actuatable properties, sampling, actuation, and the revised interpretation of observation.
- Implementation evidence: Implementations model samples, oceanographic time series, and meteorological-station measurements using SOSA’s new elements.Examples come from Geoscience Australia, the University of Bremen, and Irstea.
- Adoption: 23 ontologies and 23 datasets already reuse SOSA classes and properties.The SSN Usage document records this implementation evidence as part of the W3C standardization process.
7. Conclusion
SOSA is a lightweight ontology for describing sensors, actuators, samplers, and their activities on the Web. It supports Web-oriented interoperability through event-centric modeling, alignment with standards, and Schema.org-compatible annotations.
- SOSA represents the core of the new Semantic Sensor Network ontology and was developed jointly by the W3C and OGC.
- Its event-centric view aligns SOSA with PROV-O and O&M standards.
- Schema.org-style domainIncludes and rangeIncludes annotations let existing Schema.org applications describe IoT devices and their capabilities.
- The Spatial Data on the Web Interest Group will promote SOSA adoption and develop a binding to W3C Thing Description.