An Adaptive Multi-Objective Green Route Recommendation System Integrating Emission and Pollution Exposure

Traffic congestion, air pollution, and carbon emissions have all increased due to rapid urbanization and increased automobile use, creating serious environmental and public health issues. Conventional navigation systems primarily optimize distance or travel time, often ignoring route-level pollution exposure and vehicle-specific emissions. This paper proposes an AI model, a personalized multi-objective green route recommendation system that jointly optimizes travel time, pollution exposure, and vehicle emissions. The framework integrates spatial Air Quality Index (AQI) observations with a city road network derived from OpenStreetMap and performs efficient AQI-to-road assignment using KD-Tree nearest-neighbor mapping. A regression-based machine learning model estimates vehicle-specific emission impact, while multi-objective edge costs enable fastest and exposure aware green route planning using shortest-path optimization. Experiments conducted on Delhi NCR demonstrate that the recommended green route achieves a measurable 6.58% reduction in AQI intake exposure compared to the fastest route in a real-world case study. The system further includes explainable visualization layers (AQI heatmaps and route overlays) and an environmental impact report to support informed and sustainable mobility decisions.