Sistema colaborativo para planejamento de rotas para cadeirantes

Abstract

Route planning is a challenging problem for urban computing that usually involves the processing of a huge amount of data and collaborative user feedback. Traditionally, route planning services are street-based, that is, even paths for a pedestrian are suggested in terms of streets. However such models are not suitable for users with certain disabilities. To address this problem, we describe in this dissertation a sidewalk-based model to accommodate the needs for a wheelchair route planning service. The model is mathematically defined as a graph, where the vertices are the city block corners and the edges are the sidewalks or crosswalks. The edge costs are derived from important accessibility features, such as distance, path inclination, and existence and maintenance conditions of curb ramps, crosswalks and sidewalks. The model has been organized such as the user feedback is considered to help updating the model when accessibility issues are detected, by wheelchair users and companions, or solved, by the department of city planning. We also present a route planning algorithm that provides a set of alternative routes based on accessibility conditions, and a shortcut recommender algorithm to support accessibility related decision making by the department of city planning. Besides, it was developed an API to provide shortest-path using sidewalks. Experiments, by using PgRouting and PostGIS with open data, are reported for a Brazilian city neighborhood in order to validate the model and the route planning service.