Research Projects

The rapid expansion of four-lane and multi lane rural highways in India necessitates a robust framework for evaluating their geometric design consistency to ensure safety and operational efficiency. Traditional highway design methods often overlook how actual driver behavior—particularly vehicle operating speed—interacts with geometric and pavement features. Current global models predominantly focus on two-lane highways or use limited spot speed data, resulting in significant gaps in applicability to India’s high-speed divided rural roads. There is thus an urgent need to develop scientifically-grounded, data-driven operating speed models tailored to Indian highway conditions. This research proposes to develop predictive operating speed (V85) models for multi-lane divided rural highways, using continuous speed profiles collected via GPS-equipped vehicles under free-flow conditions. The rationale lies in bridging the critical knowledge gap in design consistency assessment by employing high-resolution speed data, capturing real-time variations in driver behavior over extended tangent curve sections. The approach integrates geometric, pavement, and traffic factors into comprehensive models, moving beyond simplistic, linear, or spot-speed-based approaches. Scientific Objectives are to analyze how roadway geometry, pavement condition, and traffic variables influence operating speeds on tangent sections; to develop statistically validated, non-linear regression models for predicting V85 based on key influencing parameters; to evaluate model performance using multiple statistical measures and assess parameter sensitivity; and to derive geometric design consistency thresholds applicable to Indian multi-lane highways. The core hypothesis is that continuous speed profiles provide significantly better predictive power for modeling V85 than conventional spot-speed methods, especially when non-linear relationships among multiple geometric and surface variables are considered. The study will test multiple functional forms (logarithmic, inverse, exponential) to determine the best-fit models for V85 prediction. It is also hypothesized that adjacent curve dynamics, pavement roughness (IRI), and tangent grade significantly influence operating speed on straight segments, contrary to assumptions in older models. Main Experiments and Methodology will comprise of Data Collection, Geometric and Pavement Data, Traffic Parameters, Preliminary Analysis, Model Development and Validation. The proposed study will contribute both fundamentally and practically to the field of highway safety and design. Scientifically, it advances operating speed modeling by introducing dynamic, multi-factor, non-linear frameworks based on real driving behavior. Practically, it will provide validated models and safety thresholds that can be directly applied to improve geometric design consistency in highway planning and reduce crash risks.