Lane-Level Localization and Map Matching for Advanced Connected and Automated Vehicle (CAV) Applications

As an emerging solution to numerous socio-economic and environmental issues in our contemporary transportation systems, Connected and Automated Vehicles (CAVs) have received a significant amount of attention from industry, government, and academia. A variety of CAV-based applications have been developed to decrease the frequency and severity of accidents, mitigate congestion, reduce energy consumption and pollutant emissions, as well as enhance system resilience and efficiency. The vast majority of such applications require accurate and reliable vehicle localization. For instance, lane-level positioning accuracy is the minimum requirement for multiple CAVs to enable Cooperative Adaptive Cruise Control (CACC) to form a platoon or to perform Cooperative Ramp Merging (CRM) to determine the most efficient merging sequence. As another example, most existing traffic signal controllers rely on fixed-location embedded loop detectors, which are usually configured to provide coarse traffic information such as the number of vehicles passing a fixed point or to detect the presence of at least one vehicle in the approach lane to an intersection. Alternatively, CAVs with high positioning accuracy can serve as reliable sensors to enable dynamic lane-level queue inventories, allowing much more effective traffic management. In addition, for many eco-ITS applications, such as eco-routing navigation and Eco-Approach and Departure (EAD), lane-level positioning accuracy would help significantly improve their effectiveness, by better estimating the driving lane, road-grade (thus the energy consumption), and the lane-level traffic information (e.g., signal phase and timing, queue length).

Due to rapid decrease in the cost and size of sensors, vehicle position accuracy and reliability is rapidly increasing. Accuracy and reliability is typically achieved by the Global Navigation Satellite Systems (GNSS) in conjunction with other sensors. Several GNSS -- e.g., GPS, GLONASS, Galileo, BeiDou -- are available, with each individually supplying 6-10 measurements per epoch, yielding a total of 24-40 single-frequency measurements, while only four (with suitable geometry) are required. In the near future, the evolution of GNSS towards multiple signals per satellite will bring the total number of GNSS signals to over 100. This increasing number of GNSS signals, when combined with differential corrections, brings the prospect of reliable lane-level (or better) vehicle position accuracy to be near at hand. This project will investigate and demonstrate the utility of lane-level localization accuracy and map-matching in a selected CAV application.

Research Area