Real-time adaptive traffic signal control systems require costly detector information. If limited traffic detectors could be used more efficiently, then implementation of adaptive signal control system would be cheaper. We can now access real time traffic flow data remotely. Sparse detector data can be manipulated to provide comprehensive real-time information. This research demonstrates how a new model, Turning Movement Estimation in Real Time (TMERT) infers unknown flows from those detected. The evaluation of the effectiveness of the quality of the estimates is comprehensive: 7,000 simulations and real-world network flow information on four separate networks with over 150 hours of traffic flow information. The paper details the relationship between actual detector data and the quality of the estimates. A location-allocation algorithm allows each link in a network to be ranked for prioritizing the detection locations. The multi-variant exponential relationship combines link flow and relative location within the network. This Utility Function’s purpose, therefore, is to enable the user to identify where to locate new permanent detectors to maximize model effectiveness. The findings identify that the TMERT model must operate with at least 30% coverage to provide reliable estimates and that it operates more efficiently at higher congestion states.
