Travel time prediction with bluetooth sensor data in intelligent traffic system (ITS)

For traffic density management, travel time plays an important role. One of the methods that can determine this period is Bluetooth technology. Bluetooth technology is an inexpensive and easy method of data collection for traffic studies. With the Bluetooth data collected by this method, information such as traffic monitoring, the ability to identify vehicles on a certain route and travel time can be obtained. It provides tracking of movements through the detection of MAC addresses specific to the Bluetooth device. With the help of Bluetooth technology, certain characteristics affecting the travel time data have been analyzed. Currently, highway travel time can be used as a new and effective data collection tool through actively used Bluetooth sensors. The central control software system includes a holistic system for collecting, formatting data in a centralized location, processing data in vehicles and providing it to drivers. Central system design, data integration, through the process from a number of sources, for example, data from sensors connected to the system, which is again in line with the related traffic information on Highway message signs (VMS) defined scenarios, text message and images, including engorged way to present the data related to the driver can be used. Providing travel time distribution information, both average and variance, can play a more effective role in ensuring that drivers are more likely to arrive on time and choosing reliable roads. For these reasons, a heterogeneous data fusion method is proposed by combining heterogeneous data from point and december detectors in order to determine the travel time distribution. Connection travel time distributions are first determined from point detector observations. It is successfully used in traffic management on highways with limited access corridors, such as highway corridors. Within the scope of this study, three case studies were developed; 1) Estimated travel time/speed along the corridors 2) Origin-Destination (OD) matrix estimation 3) Providing electronic speed control (EDS) using Bluetooth data from multiple Bluetooth readers that are activated simultaneously. The results showed that, as expected from the entrance-exit network structure of urban corridors, many MAC addresses are observed in one place at the same time by being collected in another observation location. MAC aggregate data suggest up to 10% penetration rate for Bluetooth-enabled devices in Istanbul traffic. Travel time estimation for an urban corridor was very effective using Bluetooth data, despite few MAC pairings. Moreover, the Bluetooth data provided the expected acceleration behavior before and after the EDS. The OD estimation for an open network is unreliable due to the low penetration rate of Bluetooth and the limited number of Bluetooth readers. The OD estimation process will provide greater reliability and benefits through repeated data collection and other data sources.