Delay margin comparison in a velocity-only versus headway-only connected vehicle model

Abstract

Two different connected vehicle models are considered: one based on only using velocity measurements and the other using only headway, where each model is affected by delays due to human reaction times and sensing and communication lines. The presence of delays is established between any two vehicles that are neighbors as determined by the underlying network topology. The main focus here is to put light into our understanding of how much of time delay can the equilibrium of these models tolerate before becoming unstable. To this end, we compute, using the parallel processing computational tool Delay Margin Finder (parDMF), the delay margin of the equilibrium dynamics and discuss design trade-offs. We report that delay margin in the headway-only model is much less sensitive against the presence/absence of links in the network and that this model can have larger delay margin when larger number of vehicles are capable of communicating with other vehicles in the platoon.