Sustainable urban mobility has become a critical challenge as vehicular emissions continue to be a primary driver of air pollution and climate change. This paper presents GreenHalo Maps, a lightweight, real-time eco-navigation system that integrates machine learning-based carbon emission prediction with Ant Colony Optimization (ACO) to recommend the most environmentally responsible driving route among available alternatives. Unlike conventional navigation systems that optimize solely for speed or distance, GreenHalo Maps evaluates each candidate route across five environmental parameters: trip distance, elevation gain, traffic congestion index, fuel type, and vehicle category. Two regression models — a Random Forest Regressor (250 estimators) and a LightGBM Regressor (500 estimators) — predict CO₂ emission in kilograms per route with R² scores of 0.96 and

0.95 respectively, trained on a continuously growing dataset of 1,062 real route records collected via the Google Maps Directions and Elevation APIs. An ACO algorithm operating on a coordinate graph built from decoded route polylines then identifies the minimum-emission path by optimising pheromone trails over 40 iterations with 20 ants. Air quality at each route’s midpoint is retrieved from the Google Air Quality API and surfaced as an AQI badge alongside emission figures. The system auto-retrains both models every 500 new route records logged to the training CSV, ensuring continuous model improvement without manual intervention. Deployed as a Flask web application with a Google Maps JavaScript frontend, GreenHalo Maps renders colour-coded polylines for up to three alternative routes, displays an interactive emissions comparison table, and highlights the ACO-optimised eco-route. The full pipeline installs from seven pip packages with no version conflicts across Python 3.8–3.12 and operates at interactive latency on commodity CPU hardware.

Key Terms—Eco-navigation, carbon emission estimation, ant colony optimization, random forest, LightGBM, Google Maps API, air quality index, sustainable routing, Flask, smart city.