Stewart, Connor2024-02-152024-02-152024-02-152024-02http://hdl.handle.net/10012/20344This thesis presents a comprehensive exploration of traffic rule verification systems for diverse junction types, addressing key challenges in formalizing rules, determining violation thresholds, and covering a wide spectrum of relevant traffic scenarios. Leveraging iterative implementations and extensions of existing approaches, the associated program aims to concretize literature-based methods, and understand the severity of rule violations in naturalistic driving. The study extensively tests traffic rule adherence by vehicles in simulated and recorded traffic, utilizing Lanelet2, a versatile mapping system, to cover both signalized and non-signalized stop-regulated intersections. Through statistical analyses, the research delivers results on rule-violation thresholds, associated coefficients, and traffic violation rates, encompassing scenarios such as stop sign compliance, turns after stops, traffic light violations, offroad occurrences, speed limit violations, and tailgating instances. The thesis contributes specific test cases and insights from naturalistic driving, showcasing parameter settings and threshold determination for effective traffic rule implementation. The comprehensive approach taken in this research contributes to the advancement of traffic rule verification systems and provides a foundation for evaluating autonomous vehicle behaviours in diverse junction scenarios.enArtificial IntelligenceTraffic RulesValidationknowledge representation automated reasoningMaster Thesis