Bandwidth-Efficient Marine Condition Monitoring Using Edge Analytics and Anomaly-Triggered Satellite Transmission

Zenodo metadata description This repository contains the processed dataset and reproducibility files associated with the study on bandwidth-efficient marine condition monitoring using onboard edge analytics and anomaly-triggered satellite transmission. The data were derived from a marine machinery monitoring deployment and include processed anomaly scores, transmission-policy records, burst-event indices, operational context logs, temperature and RPM time-series, summary results, and the Python script used to generate the manuscript figures. The repository is intended to support transparency, reproducibility, and secondary analysis of the reported monitoring framework. Rather than continuously transmitting high-frequency raw data over bandwidth-limited satellite links, the proposed system performs onboard processing and transmits low-rate summaries during normal operation, with selective escalation to higher-resolution data transmission only when persistent abnormal conditions are detected. The archived files document the processed outputs used to evaluate that workflow and to prepare the results presented in the manuscript. The dataset includes:(1) processed anomaly-scoring outputs used for event detection;(2) bandwidth-policy and burst-index files used to represent transmission logic and flagged anomaly intervals;(3) RPM and temperature logs used as operational monitoring variables;(4) vessel operational log data used to provide contextual interpretation of monitored events;(5) summary result tables; and(6) the script and figure folder used to reproduce manuscript visualizations. This repository does not serve as a raw sensor archive; instead, it provides the processed analytical products and reproducibility materials underlying the published study. It is suitable for readers seeking to understand, verify, or extend the event-triggered condition-monitoring and communication-reduction framework presented in the article.