Stochastic modeling for vehicle platoons (I): Dynamic grouping behavior and online platoon recognition

•A stochastic dynamic model for vehicle platoons is developed.•A Markov regime-switching stochastic process is used to model the dynamic behaviour of platoon-to-platoon transitions.•A state space model is employed to describe individual vehicles' dynamic movements within each vehicle platoon.•An algorithm is proposed for online platoon recognition.•A new method is developed for vehicle speed estimation for traffic with platoons.

A vehicle platoon is a group of vehicles traveling together at approximately the same speed. Traffic platooning is an important phenomenon that can substantially increase the capacity of roads. This two-part paper presents a new approach to stochastic dynamic modeling for vehicle platoons. In part I, we develop a vehicle platoon model with two interconnected components: a Markov regime-switching stochastic process that is used to model the dynamic behavior of platoon-to-platoon transitions, and a state space model that is employed to describe individual vehicles' dynamic movements within each vehicle platoon. On the basis of the developed stochastic dynamic model, we then develop an algorithm for online platoon recognition. The proposed stochastic dynamic model for vehicle platoons also provides a new approach to vehicle speed filtering for traffic with a platoon structure.