ADS safety-testing ontology

ADS safety-testing ontology is an OWL ontology he built during his joint research at JAIST (the Japan Advanced Institute of Science and Technology) in 2025. It structures the knowledge used to test the safety of automated driving systems (ADS). He worked with an international research team, communicating results in English.

Why an ontology

Testing an automated vehicle means covering a huge space of situations, such as road layouts, other road users, weather and manoeuvres. An ontology gives those concepts precise definitions and relationships. Test scenarios can then be described consistently, checked for gaps and queried, instead of living in free text.

Related work

• The semantic extraction pipeline extracted facts for the same testing work and validated them with SPARQL.
• The driving scenario generator explored producing scenarios with a language model.
• Graph Laboratory explores system-of-systems graphs, including automated driving.

See also

• Japan Advanced Institute of Science and Technology (JAIST)
• Ontology engineering
• Ontology and knowledge graphs