Comparative Analysis of Machine Learning Models for Evaluating the Impact of Speed Bumps on Travel Time and Fuel Consumption

Vehicle travel time and fuel consumption are greatly impacted by road surface conditions, especially the presence of speed bumps. This study investigates the effect of speed bumps on transportation efficiency using a combination of statistical analysis and machine learning approaches. The impact of various bump shapes on vehicle acceleration, deceleration, and idling, which in turn affect fuel efficiency and trip duration, is examined using empirical data gathered from controlled field studies. The actual travel time (Mean (M) = 6575.61 sec, Standard Deviation (SD) = 344.51 sec) was significantly longer than the projected travel time (M = 3689.79 sec, SD = 271.14 sec), according to a paired-samples t-test (t(432) = 145.89, p < 0.001), showing considerable delays brought on by repeated braking and acceleration. The inefficiencies caused by speed bumps were further highlighted by the much higher fuel usage than anticipated (M = 11.38 L, SD = 1.53 L vs. M = 5.12 L, SD = 0.38 L), t(432) = 111.25, p < 0.001.Based on route features, a machine learning-based predictive model was created to forecast travel time and fuel consumption. The random forest regression model had the highest accuracy (R2 = 0.94 for fuel consumption and R2 = 0.88 for travel time). According to feature importance analysis, the most significant criteria were vehicle speed, road type, and bump frequency. Furthermore, correlation analysis demonstrated a very strong correlation for fuel consumption (r = 0.962, p < 0.001) and a small but significant correlation between actual and projected travel time (r = 0.122, p = 0.011), so confirming the model's predictive dependability