This study was conducted as part of my graduation project and represents my individual contributions within the scope of the European Union–funded project "OPEVA: Optimization of Electric Vehicle Autonomy", carried out under academic supervision. As part of my graduation project, I contributed to two modules of the system: Vehicle Tracking Module and Performance Monitoring Module, providing comprehensive solutions for efficient EV fleet management.
- Real-time Monitoring: Vehicle location, speed, energy consumption, and remaining range.
- Battery Management: Continuous monitoring with critical alerts.
- Route & Delivery Tracking: Rota ilerleme ve şarj ihtiyacı analizi.
- AI-Powered Predictions:
- Segment-based energy model (100 m segments with speed, load, slope).
- CatBoost model (R² ≈ 0.94) chosen among 23 models.
- SHAP analysis for interpretability (impact of slope, speed, mass, etc.).
- Alert System: Root cause classification (vehicle, route, driver, delivery, system).
- Simulation Integration: SUMO-based simulation.
- Leaflet Map: Live visualization of vehicles and delivery routes.
- Comprehensive Analytics: 14 reports, 50+ charts.
- Time-based Analysis: Daily, weekly, monthly, yearly.
- Custom Filtering: Flexible date & time filters.
- Grafana Dashboards: Interactive charts, embedded via iframes.
- Dynamic Layouts: Accordion-based report grouping, responsive grid system.
- Multi-dimensional Reports: Vehicle, driver, route, customer, charging station, system.
- Backend: Python 3.11.9, Node.js 18.20.2
- Frontend: React (modern UI)
- Databases:
- MariaDB 11.3.2 → structured data
- MongoDB → logs & telemetry
- Visualization: Grafana dashboards (Docker-based deployment)
- Simulation: SUMO integration
- ML/Analytics: Pandas, NumPy, CatBoost, SHAP
- Processing Layer: Energy & range prediction models (CatBoost).
- Storage Layer: MariaDB + MongoDB.
- Visualization Layer: Grafana (iframe embedded) + Leaflet map.
