Empowering users in minimizing air pollution exposure during travel: a scalable algorithmic solution

This paper introduces a unique approach for health-optimal route planning in urban environments, integrating real-time Air Quality Index (AQI) and traffic congestion delay metrics into a strategically designed road network graph. The nodes represent AQI values, and the edges convey traffic congestion metrics, creating a holistic urban environment representation. User-defined weightings serve as heuristic functions for the A* algorithm, guiding the system in identifying the optimal routes. Experimental results are further validated in two areas that are densely traffic-driven, and the results affirm the efficacy of the approach. The user-driven approach provides personalized and context-aware route recommendations that balance environmental concerns and travel time. In an alternative approach, the paper explores Graph Neural Networks (GNNs) to learn improved graph embeddings before path traversals. However, compared to the previous approach, this method proves less scalable and time-optimal due to repetitive aggregation in dynamic environments. This study significantly contributes to the development of intelligent transportation systems, offering a nuanced consideration of both environmental and congestion factors for enhanced urban route planning.