Road-Speed Profile for Enhanced Perception of Traffic Conditions in a Partially Connected Vehicle Environment

Anticipative driving of vehicles, considering the microscopic dynamics of the preceding traffic, can bring great improvement in the traffic flows in a connected vehicle environment (CVE), where 100% vehicles frequently share their states using vehicle-to-vehicle communication technology. This paper addresses a partially CVE, where unconnected (manually driven) vehicles are mixed in the traffic, and proposes a novel method of comprehending traffic conditions that can be used for realizing highly anticipative driving. More specifically, we refer to anticipative driving as the predictive control of a host vehicle considering its preceding traffic conditions in an extended view. For enhanced perception of the traffic conditions on the road, a road-speed profile is proposed that concisely describes the mean speed in each small segment or cell of the road by effectively extracting information from traffic big data, i.e., broadcasted data from all surrounding vehicles to the host vehicle. This process of dynamically approximating the road-speed profile is performed in two steps. First, a conditional persistence prediction model is proposed to estimate the future states of the connected vehicles. Such predicted states can sufficiently compensate missing data from any unconnected vehicles. Second, the predicted time, position, and speed of the vehicles are then mapped onto the road cells, and subsequently, the corresponding cell speed is tuned by employing the weighted-moving average technique. Accuracy of both the persistence prediction model and the road-speed profile are empirically evaluated for different penetration of the connected vehicles. Finally, an anticipative driving scheme in a model predictive control framework is developed by incorporating the proposed road-speed profile, and performance of the anticipative driving is compared with the existing scheme that needs the CVE.