String Stability of Connected Vehicle Platoons under Lossy V2V Communication

TL;DR

This paper investigates string stability in connected vehicle platoons under lossy V2V communication, focusing on safety-related error bounds and optimal headway settings to prevent collisions.

Contribution

It redefines string stability from a safety perspective, providing bounds on spacing errors and minimum headways under packet loss conditions, validated through vehicle modeling and experiments.

Findings

Upper limit on maximum spacing error based on lead vehicle acceleration

Minimum achievable time headway under burst-noise packet losses

Validated vehicle model matching Lincoln MKZ performance

Abstract

Recent advances in vehicle connectivity have allowed formation of autonomous vehicle platoons for improved mobility and traffic throughput. In order to avoid a pile-up in such platoons, it is important to ensure platoon (string) stability, which is the focus of this work. As per conventional definition of string stability, the power (2-norm) of the spacing error signals should not amplify downstream in a platoon. But in practice, it is the infinity-norm of the spacing error signal that dictates whether a collision occurs. We address this discrepancy in the first part of our work, where we reconsider string stability from a safety perspective and develop an upper limit on the maximum spacing error in a homogeneous platoon as a function of the acceleration maneuver of the lead vehicle. In the second part of this paper, we extend our previous results by providing the minimum achievable time headway for platoons with two-predecessor lookup schemes experiencing burst-noise packet losses. Finally, we utilize throttle and brake maps to develop a longitudinal vehicle model and validate it against a Lincoln MKZ which is then used for numerical corroboration of the proposed time headway selection algorithms.