An agent-based inter-vehicle cooperative robust car-following model for longitudinal control under uncertainty

Abstract

Uncertainty is common challenge facing system modelling and simulation in uncertain environments where the outcome of control actions is unknown. In this paper, an agent-based robust car-following model is proposed to handle uncertainty in the longitudinal control system based on inter-vehicle cooperation where an organisation of heterogeneous agents communicate and cooperate in a vehicular ad hoc network to provide reliable and collision-free control policies. We used agent-based modelling to model the inter-vehicle cooperative longitudinal control system while integrating human drivers in the multi-agent system to assume specific roles and cooperate with other agents to accomplish the task at hand. For behavioural analysis and comparison with other car-following models, microscopic agent-based simulations conducted using JADE and SUMO showed that the proposed model offers collision-free control of vehicle speed under different conditions including uncertainty. Also, we used the proposed model to investigate the effect of uncertainty on both vehicle's speed and its control actions.