Analysing and Modelling the Impact of Aggressive Driving on Crash Accident Duration: Based on Telematics Data Driven

Abstract

Aggressive Driving Behaviours Such As Overspeeding, Harsh Braking, And Erratic Steering Are Widely Recognized As A Major Cause Of Road Traffic Crashes, Yet Limited Research Has Focused On Modelling The Time To Crash Based On These Behaviours Using Real-World Telematics Data. This Study Examine How These Behaviours Affect The Time Until A Crash Occurs, Using Six Months Of Pre-Crash Telematics Data From 43 Cross-Country Public Buses In Ethiopia. A Quantitative, Predictive Modelling Approach Was Employed To Systematically Examine The Relationship Between Aggressive Driving Behaviours And Timing To Crash Incidents Using Numerical Telematics Data. Survival Analyses, Cox Proportional Hazards Regression Modelling Founded That Overspeeding Is The Most Significant Predictor Of Early Crash Risk. Each Unit Increase In The Overspeeding Index Is Associated With A 17% Increase In Crash Hazard (Hr = 1.17, P < 0.001), Confirming As The Leading Risk Factor. Harsh Braking Also Contributes Significantly To Crash Likelihood, Showing An 8% Increased Hazard Per Unit Increase (Hr = 1.08, P = 0.005). Erratic Steering, Measured Through Lane Departure Frequency, Has A Moderate But Statistically Significant Effect, Increasing Crash Hazard By 6% Per Unit (Hr = 1.06, P = 0.027). Driver Behaviour Was Quantified Through Telematics-Derived Metrics Overspeeding Duration (Seconds), Harsh Braking Events, And Lane Departure Events Each Normalized Per 100 Km Of Driving Distance. Only Crash Cases Where The Psts Driver Was Identified As The Primary At-Fault Party, And The Cause Was Directly Linked To Aggressive Behaviour, Were Included In The Analysis. The Study Employed Descriptive Statistics And Correlation Analysis (Including Visualizations), Followed By Survival Analysis Using Cox Regression And Kaplan?�?Meier Estimators. All Analyses Were Conducted Using Jamovi (V2.4.5) And Jasp (V0.18.1) Open-Source Platforms Based On Spss. The Findings Confirm That Telematics Based Behavioural Monitoring Can Be Used To Build Predictive Models For Estimating Time-To-Crash, Providing An Opportunity For Early Intervention. The Study Recommends That Transport Authorities And Fleet Managers Adopt These Tools To Identify High-Risk Drivers, Initiate Timely Retraining, And Implement DataInformed Road Safety Policies.