Platoon Control Algorithm Evaluation: Metrics, Configurations, Perturbations, and Scenarios

Abstract

Vehicles are being equipped with more and more smart devices, which help the driver in his tasks. Alongside the trend to more and more autonomous vehicles emerges the possibility of making vehicles that move together as a platoon, which can be defined as a spatio-temporal organization of a set of vehicles based on a specific predetermined geometrical configuration. Basically, there are two main approaches for performing platoon control and these depend on the reference frame used (local or global). Even if the literature about platoon control is abundant, the most commonly used metrics are only tied to the stability of the platoon, that is, to the lateral and longitudinal distances between vehicles. However, these metrics are sufficient enough for a first evaluation of one algorithm; a deeper evaluation of a platoon control solution requires more well-defined metrics and perturbations that allow for the testing of algorithms with various parametrized conditions. The goal of this article is to present definitions of metrics, perturbations, spatial configurations, and scenarios that are aimed at helping scientists in the field to determine strong and precise evaluations of their platoon control algorithms. The work presented in this article stems directly from thinking performed during the institutional projects in which the authors have been involved and also takes inspiration from numerous articles on the subject and experiments made both in simulation and with real vehicles.

AVDM: A hierarchical command-and-control system architecture for cooperative autonomous vehicles in highways scenario using microscopic simulations

“…These microscopic models are, however, not sufficient to capture the entire spectrum of AV mobility since most of them do not include high-level planning and control needed for platoon formation [70,76]. Plenty of resources have been consecrated to develop platoon models featuring join and split manoeuvres [13,14,29]. However, less attention was directed to develop platoon models for microscopic simulation.…”

Modelling and simulation of a new urban freight distribution system based on automatic van platooning and fixed split up locations

“…Platoons of automated vehicles are considered, in which only the first one is driven and the others are driverless. Within this project, innovative platoon control algorithms have been developed especially suited for congested urban environments; they are described in Gechter et al [19] and Avanzini et al [20].…”

Distributed Nonlinear Model Predictive Control and Metric Learning for Heterogeneous Vehicle Platooning with Cut-in/Cut-out Maneuvers

“…Our method is general for heterogeneous vehicles and can handle both cut-in and cut-out maneuvers while tracking the desired speed trajectory and ensuring the safe desired gap between any two consecutive vehicles. Secondly, we analyze driving experience, including driving comfort, fuel economy, and absolute and relative convergence by analyzing the metrics in DNMPC [16] using distributed metric learning [17] and Alternating Direction Method of Multipliers (ADMM) optimization [18] which is found to be useful for MPC [19].…”