Alejandro Berlanga
For many border communities, crossing on foot is part of daily life. Commuters, students and shoppers move through pedestrian facilities at land ports of entry every day, often navigating long and unpredictable wait times along the way.
Yet compared to vehicle crossings, pedestrian movement remains far less understood.
While agencies have made significant progress in measuring wait times for commercial and passenger vehicles, similar insights for pedestrians are limited. Data are often sparse, collected manually or not available in real time, making it difficult for planners and operators to fully understand how pedestrian facilities are performing or where improvements are most needed.
To help address that gap, researchers at the Center for International Intelligent Transportation Research are exploring how artificial intelligence (AI) can support a more consistent and scalable approach to measuring pedestrian border crossing times.
From observation to automation
Recent research has focused on applying computer vision and person re-identification techniques to estimate how long it takes pedestrians to move through a border crossing.
Using existing surveillance infrastructure, researchers analyzed video footage from cameras located on both sides of the border, including near the toll booth area in Mexico and at the pedestrian exit point in the United States. This binational setup allows for the tracking of pedestrian movement across multiple points in the crossing process without introducing new hardware or disrupting operations.
Rather than identifying individuals, the system relies on AI models to recognize patterns and match anonymous visual features as pedestrians move between camera views. This approach supports the estimation of travel times while maintaining privacy and avoiding the need for personal data collection.
What makes this effort distinct is its ability to work within complex and real-world conditions. Pedestrian crossings often involve irregular flows, changing queues and multiple pathways, making them more difficult to measure than structured vehicle lanes. By leveraging AI to interpret these environments, the approach offers a path toward automated monitoring that adapts to the realities of pedestrian movement.
What better data makes possible
Reliable information on pedestrian wait times has implications well beyond the crossing itself.
For agencies, consistent measurement can support more informed operational decisions, from staffing and facility management to identifying bottlenecks and evaluating improvements over time. For planners, it provides a clearer understanding of how pedestrian demand interacts with surrounding transportation systems, including transit connections and urban mobility patterns.
In regions where cross-border travel is a daily necessity, these insights can help improve the overall efficiency and accessibility of the transportation network.
The ability to generate this information using existing infrastructure also makes the approach more practical to implement. By reducing the need for new equipment and manual data collection, AI-based methods offer a lower-cost and more scalable option for expanding pedestrian monitoring across multiple crossings.
Looking ahead
This pilot represents an early step in applying AI-driven analysis to pedestrian mobility at the border.
Opportunities remain to expand the approach across additional locations, refine model performance under varying conditions and integrate pedestrian data with broader transportation system metrics. As technologies continue to evolve, combining video analytics with other data sources may further enhance the accuracy and usefulness of these insights.
Beyond border environments, similar methods could support applications such as transit station analysis, pedestrian origin-destination studies and active transportation planning in urban areas.
By improving how pedestrian movement is measured and understood, this research helps lay the groundwork for more data-driven decisions that support mobility, efficiency and quality of life in border communities.
Alejandro Berlanga is a software engineer II with the Center for International Intelligent Transportation Research (CIITR) in TTI’s El Paso Office. The project was sponsored by CIITR, with support from the City of El Paso and Fideicomiso de Puentes Fronterizos de Chihuahua. The City of El Paso assisted with identifying camera installation locations, while Fideicomiso provided access to its facilities and supported data collection efforts. The research team included Erik Vargas, Alfredo Guzman, Ayla Turner and Swapnil Samant.