Evaluating Fail-Safe Traffic Signal Control under Communication Loss: SUMO Simulation Code, Datasets, and Reproducibility Package
Abstract
This repository provides the simulation code, experimental datasets, traffic-network configurations, statistical analysis scripts, and figures supporting the research manuscript “Evaluating Fail-Safe Traffic Signal Control under Communication Loss,” prepared for submission to Transportation Research Part C: Emerging Technologies. The study investigates how communication failures and fail-safe traffic signal control influence vehicle delay, particularly when traffic-signal controllers must operate without reliable vehicle-count information. The experiments examine the interaction between fallback green-time allocation, healthy controller behavior, communication disruptions, and queue storage. Repository Contents The repository includes: Python scripts for SUMO simulation and statistical analysis. Three simulated traffic configurations: an isolated intersection, a 400-meter two-signal corridor, and an 80-meter short-link corridor. A main dataset containing 6,140 simulation runs. An additional dataset containing 240 extended-horizon simulations. Comparisons of equal-budget and flow-proportional green-time allocations. Experiments involving communication loss, detector operating-point assumptions, response delay, false alarms, and controller behavior. Paired statistical comparisons, confidence intervals, sensitivity analyses, and manuscript figures. Documentation, software requirements, citation metadata, licensing information, and SHA-256 checksums. Methods and Reproducibility The archived simulations were conducted using Eclipse SUMO 1.28.0 and Python. The controller retains the last available vehicle-count information during short communication interruptions and activates a fixed green-time fallback after five consecutive missed reports. The experiments compare alternative green splits under different traffic demands, controller policies, and communication conditions. An extended-horizon sensitivity analysis evaluates the consequences of unfinished vehicle trips in the short-link scenario. The repository includes instructions for reproducing the analysis and running the simulation experiments in a clean working directory. Important Limitations Detector decisions are simulated using specified false-alarm and missed-detection probabilities; they are not actual held-out predictions from trained cybersecurity classifiers. The archive does not include CICDDoS2019 or CICIoT2023 traffic records, trained detector checkpoints, or measured detector event streams. The communication-loss scenarios are modeled rather than calibrated against an independently measured production traffic-management communication network. The findings should therefore be interpreted within the tested simulation configurations and assumptions. Software and Licensing Simulation software: Eclipse SUMO 1.28.0. Code: MIT License. Result datasets: Creative Commons Attribution 4.0 International (CC BY 4.0). Eclipse SUMO itself is not redistributed in this repository. Related Manuscript Evaluating Fail-Safe Traffic Signal Control under Communication Loss. Target journal: Transportation Research Part C: Emerging Technologies. Repository version: 7.0.0.