V2V‐enabled cooperative adaptive cruise control strategy for improving driving safety and travel efficiency of semi‐automated vehicle fleet

Abstract

AbstractAlthough highly automated vehicle fleets have committed a safe and efficient traffic flow, the next few years are likely the human‐driven vehicles (HDV) and intelligent vehicles (IV) mixed in large‐scale traffic network, where the interaction of HDVs with IVs will become more frequent. In this study, the authors investigate the traffic flow performance in a widespread of V2V environment deployed for all types of HDVs and IVs, levelled from L0 to L5. Based on this assumption, the vehicle movements and operations are collected and assembled into a V2X BSM dataset, and to any vehicle in fleet, the sequential movements and driving behaviours of the vehicle ahead can be predicted in the short term and considered as the input to cooperative adaptive cruise control (CACC) model of the controlled vehicle. The predicted motion of the vehicle ahead is applied to rolling optimization process of a given MPC framework, which calculates appropriate longitudinal operation for each car‐following vehicle and keeps optimal equilibrium state of the vehicle fleets in heterogeneous traffic flow. The experimental results showed that when the front vehicle generates random acceleration/deceleration in saturated and over saturated traffic state, in the car‐following case, the improved CACC controller maintains average car‐following distance at 20–40 m, the acceleration variation in the range [−0.1, 0.25], the gap error changed within [−10, 13], and controls the relative speed variation from −0.25 to 0.3 m/s, compared to an average headway of 20–140 m, acceleration variation in the range [−1.6, 0.65], gap error changed within [−21, 30], and a relative speed variation from −1.75 to 4.1 m/s by the traditional CACC controller. In the car‐in and car‐out cases, the vehicle controlled by the improved CACC controller had a travel displacement of 4050 m, a speed fluctuation interval of [18.7, 27.7], and an acceleration variation range of [−0.15, 0.2], while the vehicle controlled by the traditional MPC controller had a travel displacement of 3800 m, a speed fluctuation interval of [13.8,27.7] and an acceleration variation range of [−2.9, 0.6]. The improved CACC model can adapt more accurately and quickly to the maneouvre of vehicle ahead in V2V environment, which can effectively improve the capacity of heterogeneous traffic flow and the safety and comfort of driving.