Abstract:
This research focused on the development of accident prediction models for Akure-Owo highway. These models relate accident numbers, as a dependent variable, with possible causes of accidents which could be due to human, vehicle, road and environmental factors, as independent variables. This research developed a regression model using the accident causative factors. The regression model predicts accident rate along the route (the dependent variable) based on independent variables which are the human and vehicle factors as obtained from variable correlation. The model is a strong model with R2 of 0.998. More so, a Mathematical Accident Prediction Model using Spot Speed, Condition of Shoulders, Pavement Condition, Width of Pavement, Gradient of Pavement, Intersection, Drainage Condition, Stopping Sight Distance and Overtaking Sight Distance as parameters was developed, from which the acronym SCPWGIDSO was formed. Accident data for six years (2010 – 2015) were obtained from Federal Road Safety Commission (FRSC) and used to identify accident prone locations along Akure-Owo highway. Twenty-four locations called “Black Spots” were identified. Field data such as, Speed of Vehicles, Condition of Shoulder, Pavement Condition, Width of Pavement, Gradient of Pavement, Intersection, Drainage Condition, Stopping and Overtaking Sight Distance information were obtained from road condition survey. The SCPWGIDSO-AV rating system and weights produced a calibrated mathematical model that gave impressive results when validated with field data. The result revealed that the Total SCPWGIDSO-AV Indices for ten (10) locations with more than ten times accident occurrence are Emure Uli (126), Rufus Giwa poly second gate (138), Shasha (151), Rufus Giwa poly main gate (151), Bolorunduro (157), Olufoam (172), USO (172), Ogbese (175), Roadblock (178) and NNPC mega station (184) and gave similar patterns with accident records that are two times the other 14 locations
