Pre-trip energy-use prediction for micro electric vehicles from a single input: Driving-Profile Extraction and the Single Feature Prediction Model

Accurate pre-trip forecasting of energy use in micro electric vehicles (MEVs) supports planning and mitigates range uncertainty, yet many models require multi-source inputs that are unavailable before departure. We present a data-efficient framework that predicts trip energy use from a single planned input (distance) by combining a Driving-Profile Extractor (DPE) with a multi-feature predictor into the Single Feature Prediction Model (SFPM). The DPE converts planned distance into profile-based features and blends dual driving profiles through a distance-dependent weight, enabling personalized forecasts without real-time signals. We further quantify uncertainty via bootstrap prediction intervals, and derive two actionable outputs: available driving range for a given battery state and trip success probability at the planning stage. Using real-world MEV logs across multiple vehicle models and trip-distance regimes, SFPM captures vehicle-specific driving characteristics and delivers stable, high-accuracy predictions with calibrated intervals. The results indicate that a single-input, profile-generation approach can provide practical, reproducible, and uncertainty-aware forecasts suited to pre-trip decisions and subsequent analysis. The framework is readily extensible to fleet-level applications and integration with charging or routing tools.