Average speed is crucial for calculating link travel time to find the fastest path in a road network. However, readily available data sources like OpenStreetMap (OSM) often lack information about the average speed of a road. However, OSM contains other road information which enables an estimation of average speed in rural regions. In this paper, we develop a Fuzzy Framework for Speed Estimation (Fuzzy-FSE) that employs fuzzy control to estimate average speed based on the parameters road class, road slope, road surface and link length. The OSM road network and, optionally, a digital elevation model (DEM) serve as free-to-use and worldwide available input data. The Fuzzy-FSE consists of two parts: (a) a rule and knowledge base which decides on the output membership functions and (b) multiple Fuzzy Control Systems which calculate the output average speeds. The Fuzzy-FSE is applied exemplary and evaluated for the BioBío and Maule region in central Chile and for the north of New South Wales in Australia. Results demonstrate that, even using only OSM data, the Fuzzy-FSE performs better than existing methods such as fixed speed profiles. Compared to these methods, the Fuzzy-FSE improves the speed estimation between 2 % to 12 %. In future work, we will investigate the potential of data-driven machine learning methods to estimate average speed. The applied datasets and the source code of the Fuzzy-FSE are available via GitHub.

No abstract

No abstract

Average speed information, which is essential for routing applications, is often missing in the freely available OpenStreetMap (OSM) road network. In this contribution, we propose an estimation framework, including different machine learning (ML) models that estimate rural roads’ average speed based on current road information in OSM. We rely on three datasets covering two regions in Chile and Australia. Google Directions API data serves as reference data. An appropriate estimation framework is presented, which involves supervised ML models, unsupervised clustering, and dimensionality reduction to generate new input features. The regression performance of each model with different input feature modes is evaluated on each dataset. The best performing model results in a coefficient of determination R2=80.43%, which is significantly better than previous approaches relying on domain-knowledge. Overall, the potential of the ML-based estimation framework to estimate the average speed with OSM road network data is demonstrated. This ML-based approach is data-driven and does not require any domain knowledge. In the future, we intend to focus on the generalization ability of the estimation framework concerning its application in different regions worldwide. The implementation of our estimation framework for an exemplary dataset is provided on GitHub.

No abstract