Automatic Real-Time Detection and Correction of Erroneous Detector Data with Fourier Transforms for Online Traffic Control Architectures

Real-time and Internet-based control architectures are currently being designed in the context of online route guidance and traveler information systems for vehicular traffic networks. They involve transmitting field data to the traffic control center for real-time processing. To enable reliable and uninterrupted operation, these real-time systems should be fault tolerant to critical hardware failure modes such as malfunctioning detectors and failed transmission and communications links. This research proposes a Fourier transform-based fault-tolerant framework to detect data faults due to malfunctioning detectors and predicts the likely actual data for seamless operation of an online traffic control architecture. However, incidents also exhibit data characteristics similar to those of some hardware-related data faults. The proposed approach treats data faults and incidents as abnormalities in the monitored network. It first detects an abnormality and then distinguishes data faults from incidents by using a Fourier transform-based approach. Data faults are corrected with a Fourier transform-based data correction heuristic. Field data from Athens, Greece, and from Hayward, California, are used to validate the proposed methodologies. The approach uses data directly without elaborate modeling, circumventing likely modeling errors and enabling simpler adaptability to future demand–supply changes. A key contribution is the ability to robustly predict near-term traffic conditions efficiently by using historical data and immediate past data on the current day. This alleviates the computational burden and enables minute-to-minute traffic prediction, substantially aiding the seamless and uninterrupted operation of online traffic control architectures.