Fuzzy Logic Based Automation of the Extraction of Surrogate Safety Measures and the Creation of Severity Classification Using Video Data
摘要
This paper uses fuzzy logic to create an Artificial lntelligence (A.I.) based automated system that mimics expert rating of traffic conflicts. Video data from multiple sites were collected to study traffic conflicts as surrogate safety measures for traffic collisions. As part of that effort human trained subjects were given instructions to analyze traffic conflicts and assign severity levels to those. This paper proposes a fuzzy logic-based A.I. system that is trained based on such data so that the process of severity assignment can be done by the software system.