Ontology Modeling of the Estonian Traffic Act for Self-driving Buses

• Published in: Information Management and Big Data Vol. 898; pp. 249 - 256
• Main Authors: Nogales, Alberto, Täks, Ermo, Taveter, Kuldar
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2019; Springer International Publishing
• Series: Communications in Computer and Information Science
• Subjects: Deep learning; Ontology learning; Ontology matching
• ISBN: 9783030116798; 3030116794
• ISSN: 1865-0929; 1865-0937
• DOI: 10.1007/978-3-030-11680-4_24

The development of self-driving cars is a major research area that has led to several still unresolved issues. One of them is the need to abide by the legal stipulations fixed by a traffic act concerning the territory of operation. An appropriate solution to make text understandable by machines is the use of ontologies. This paper presents a first approach where the Estonian Traffic Act is transformed from text into populated ontologies, so it can be understood by machines. The proposal is a (semi)-automatic ontology learning process that combines natural language processing (NLP) and ontology matching techniques with a deep learning model. The results show that 78% of the norms that have been considered valid can be modelled with the method described in the paper.