Abstract
Traffic state estimation (TSE) is essential for road traffic management. However, TSE covering wide spatial domain is practically difficult due to the technical limitations regarding data collection. To ll this gap, this study proposes a TSE method for wide spatial domain using remote sensing by small satellites and mobile sensing by probe vehicles. Specically, the remote sensing is supposed to measure traffic density in the entire road network with few hours interval, whereas the mobile sensing is supposed to measure traffic speed continuously. The proposed method estimates the traffic state by combining these measurements by using data assimilation based on traffic ow theory. The notable feature of the method is that it endogenously estimates fundamental diagram parameters and penetration rate of probe vehicles, meaning that the method is completely free from detectors. The performance of the proposed method was evaluated by numerical simulation. It was found that few hours time interval of remote sensing and 10% or more penetration rate of probe vehicles were required for accurate estimation.
