Simulating longitudinal vehicle dynamics in heterogeneous and disordered traffic: a Non-Lane-Based Intelligent Driver Model

This study presents a comprehensive modeling framework for simulating the longitudinal movement of vehicles in heterogeneous and disordered traffic (HDT) conditions, typical of many developing countries. Traditional car-following models, designed for structured, lane-based traffic environments, often fall short in capturing complex behavioural dynamics and diverse vehicle interactions characteristic of HDT. To address these limitations, we introduce the Non-Lane-Based Intelligent Driver Model (NLBIDM), an extension of the classical Intelligent Driver Model, incorporating four key innovations: minimum lateral clearance, dynamic pseudo lane, centerline separation ratio and vehicle-specific parameters. These enhancements enable adaptive leader identification, realistic modeling of staggered following behaviour, and the representation of lateral–longitudinal movement interactions, while accounting for the distinct physical and behavioural characteristics of a heterogeneous vehicle fleet. Steady-state conditions are derived analytically, and a stability assessment is carried out to deepen theoretical insight into equilibrium dynamics. The model is calibrated using high-resolution trajectory data collected in Chennai, India − an environment representative of HDT conditions. Calibration results produce realistic and interpretable parameter estimates across various vehicle classes, with root mean square normalized errors falling within acceptable ranges. Validation under multiple parameter assignment scenarios further confirms the robustness and generalizability of the proposed model. Overall, the NLBIDM provides a behaviourally enriched and versatile framework for microscopic traffic simulation, with significant potential for application in traffic management and transport planning.