DiPAS: A Digital Pollen Avoidance System as an Innovative Solution for Hay Fever Management

Hay fever is a globally prevalent but underserved chronic disease. Despite previous research into traditional and digital hay fever management, there is a lack of personalised guidance for its prevention. The aim of this research is to develop a digital pollen avoidance system that provides individualised support for symptom mitigation and prevention. To this end, we proposed DiPAS, a system that minimises outdoor pollen inhalation by generating data-driven optimal walking routes. This is achieved through three stages: training segmentation neural networks to identify grass regions, developing route optimisation pipeline using Generalised Voronoi Diagram (GVD) and A* search algorithm, and adopting the Cellular Automaton Model (CAM) to account for wind effects on pollen distribution. DiPAS achieved promising segmentation performances, with a Dice score of 0.81 and an IoU of 0.71. The subsequent route optimisation and CAM were validated with both simulated and real-world samples quantitatively and qualitatively, demonstrating accurate path planning and adaptive adjustments under varying conditions. A final real-world case study further illustrated the system’s full end-to-end workflow and practical applicability. Overall, DiPAS shows effectiveness as a digital solution to hay fever management, showcasing its potential as a personalised healthcare tool to improve patients’ quality of life. Our implementation is available at https://github.com/yuhanzheng234/DiPAS.git.