Intelligent Car Crash Reporting System

Abstract

Malaysia is one of the most traffic accident vulnerable countries in the world. This is due to the reason of high reliance on vehicles and roads as the major transportation in the Malaysian context. The gradual increase in road accidents has caused the congestion of Malaysian Emergency Response Services (MERS) which has further led to problems related to the failure of timely handling bulk of cases thus inadvertently causing fatalities. To mitigate the issues related to the inefficient emergency response measures, an intelligent car crash reporting system known as Autoport has been proposed. The aim is to streamline the process of car crash reporting, diverge the call traffic from the MERS hotline, as well as, reduce the prank calls by implementing machine learning and object detection for automating the reporting task. Once a user uploads a video for reporting, the proposed solution will detect the severity of the car crash and send rescue notification to required authorities via push notifications. The proposed solution is developed with the adoption of the Rapid Application Development methodology. The paper also highlights the system requirements gathering processes by the means of interviews and questionnaires. Besides, the accident severity detection model development and evaluation played a pivotal role in enhancing the validity of the work. This research also includes the analysis of User Acceptance Testing, further validates the proof of concept and viability of the proposed solution in streamlining and fastening the car crash reporting process for better rescue experiences.