Exploring hybrid models for identifying locations for active mobility pathways using real-time spatial Delphi and GANs

Abstract

The spatial planning process is considered an extremely complex system, as it comprises different variables that interrelate and interact with each other. Effectively addressing this spatial complexity necessitates a multidisciplinary approach, as unified methodologies may prove insufficient. Specifically, in urban planning, it is increasingly crucial to prioritize bike lanes, bike stations, and pedestrian zones, for functional transportation infrastructures. This approach can enhance cities by improving air quality, reducing emissions, and boosting public health and safety through physical activity and accident prevention. However, implementing these changes requires careful planning, community engagement, and stakeholder collaboration. This paper proposes a hybrid model for identifying optimal locations for bike lanes, bike stations, and pedestrian zones adopting Real-Time Spatial Delphi and Generative Adversarial Networks (GANs). The Real-Time Spatial Delphi is a modified version of the traditional Delphi method that incorporates real-time feedback and visualization of group response in real-time, aiming to achieve a convergence of opinions among experts on the territory. Nevertheless, these judgments are a spatial representation not visible in reality, and with the spread of artificial intelligence models, different implementations can support the planning process, such as the use of GANs. In this case, GANs can be exploited by adopting pre-existing location images resulting from experts' judgments to illustrate the proposed intervention's visual impact. To demonstrate the effectiveness of our hybrid model, we apply it to the city of Dublin. The results showcased how the method helps stakeholders, policymakers, and citizens in visualizing the proposed changes and gauging their potential impact with greater precision.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.