Road Maintenance Prioritization (RMP) System

Abstract. This study develops a transparent, weighted framework for assessing road condition and prioritizing maintenance at the segment level. Traffic data (AADT/ESAL) are fused with climate and terrain layers extracted via Google Earth Engine (GEE) into a national spatial database. A Final Road Condition Index (FRCI; 0–100) is derived from eight weighted criteria (CTDI, slope, vegetation, surface water, topographic wetness, precipitation, snow, erosion), updated periodically to reflect changing conditions. The framework was applied to Türkiye’s national road network managed by the General Directorate of Highways (KGM). Environmental rasters were sampled every 5 km along road centerlines, traffic was converted to ESAL and normalized to CTDI, and the FRCI was computed as a weighted sum. An interactive Streamlit dashboard with a MySQL/PostgreSQL backend enables visualization, sensitivity testing, and AI-driven treatment recommendations, benchmarked against KGM’s current inspection-based system. Results show that incorporating GEE layers improved prioritization compared to AADT-only baselines, increased decision consistency, reduced time-to-decision, and yielded higher benefit under fixed budgets. The suggestion engine provided more consistent, better-justified recommendations. The study recommends adopting the FRCI framework nationally, institutionalizing periodic GEE updates, formal governance of weights and criteria, and embedding the dashboard and suggestion engine into KGM’s annual planning cycle for more efficient, evidence-based maintenance.