Trajectory reconstruction for freeway traffic mixed with human-driven vehicles and connected and automated vehicles

The development of technologies related to connected and automated vehicles (CAVs) allows for a new approach to collect vehicle trajectory. However, trajectory data collected in this way represent only sampled traffic flow owing to low penetration rates (PRs) of CAVs and privacy concerns, and fail to provide a comprehensive picture of traffic flow. This study proposes a method to reconstruct vehicle trajectories in fully sampled traffic flow on freeways that consists of human-driven vehicles (HVs) and CAVs by using the mobile sensing data acquired from CAVs. The expected behavior of vehicles within the detection range of CAVs is determined based on the driving state classified by the Wiedemann model, i.e., free driving, emergency, closing, and following. If the actual behavior is different from the expected, it is deemed to be influenced by the undetected HVs. Then, new HVs are inserted based on the estimated local traffic density and speed of the freeway. The trajectories of the inserted HVs are further reconstructed by using the established update rules of cellular automation, i.e., uniform motion, acceleration, deceleration, randomization and position update. Last, the proposed method was examined by simulation experiments under different traffic densities and PRs of CAVs. The results show that the trajectories of fully sampled mixed traffic flow can be reconstructed reasonably well, not only under traffic conditions without explicit congestion but in congested environments.