A Collaborative System for Suitable Wheelchair Route Planning
Guilherme L. Barczyszyn, Letícia M. De O. Camenar, Diego De F. Do Nascimento, Nádia P. Kozievitch, Ricardo D. Da Silva, Leonelo D. A. Almeida, Juliana De Santi, Rodrigo Minetto · 2018 · ACM Transactions on Accessible Computing · doi:10.1145/3237186
Summary
This paper presents a collaborative system for wheelchair route planning that fundamentally reimagines how navigation services work for people with mobility impairments. Unlike standard route planners like Google Maps that provide street-based directions, this system uses a sidewalk-based model where routes are defined in terms of individual sidewalks and crosswalks rather than streets. The research began with extensive requirement analysis involving wheelchair users in Curitiba, Brazil, including 29 questionnaires, 12 semi-structured interviews, focus groups with 16 participants, and direct observation sessions. Participants consistently reported that sidewalk conditions—not just distances—determine whether routes are actually usable. Issues like missing curb ramps, steep slopes, poor pavement conditions, and obstructions were identified as critical barriers. The system is built on a graph model where vertices represent city block corners and edges represent sidewalks or crosswalks. Edge costs incorporate multiple accessibility factors: distance, path inclination (calculated from topographic contour data), and an accessibility score reflecting the presence and condition of curb ramps, sidewalk quality, and obstructions. The system was implemented using PostGIS and PgRouting for geographic processing. A key innovation is the collaborative feedback loop: wheelchair users can report accessibility issues through a mobile app, city planning departments can mark problems as resolved, and the routing model updates accordingly. This creates a living map that reflects actual on-the-ground conditions rather than idealized infrastructure.
Key findings
The system was validated in the Batel district of Curitiba, a neighborhood with 1.76 square kilometers and 46 streets. Testing revealed that only 33% of households in Curitiba have sidewalks and just 12.6% have curb ramps—even in this relatively wealthy area, only 53% of required infrastructure exists. The routing algorithm provides multiple path alternatives optimized for different criteria: shortest distance, minimum inclination (staying below the 8% grade limit specified by Brazilian accessibility standard NBR9050), and fewest accessibility problems. In one test, the distance-optimized path was 907.47m with 13.05% maximum slope effort, while the inclination-optimized path was 1,021.91m but kept slopes below 6.99%. The shortcut recommender algorithm demonstrated that strategically adding just 20 curb ramps could reduce average path distances from 600m to approximately 30m—showing city planners where limited infrastructure budgets would have the greatest impact on wheelchair mobility. User evaluations found participants would use the mobile app daily, though they needed tutorials to understand the interface icons.
Relevance
This research offers a practical template for building inclusive navigation services. The sidewalk-based graph model could be adapted to any city using open GIS data, and the collaborative feedback mechanism addresses the reality that accessibility conditions change constantly due to construction, weather damage, and repairs. For practitioners, the key insight is that route planning for wheelchair users requires granular, segment-level accessibility data—not just whether a route exists but whether each sidewalk and crossing is actually usable. The shortcut recommender provides an evidence-based approach for advocacy: showing city officials exactly which infrastructure improvements would have the greatest impact. The research also highlights that wheelchair users often prefer longer accessible routes over shorter inaccessible ones—a finding that should inform how navigation apps present options to users with mobility impairments.
Tags: mobility accessibility · wheelchair users · route planning · crowdsourcing · geographic information systems · urban computing · participatory design
Standards referenced: NBR9050 · ADA