Click here to enter text. Characterizing parking systems from sensors data through a data-driven approach Jamie Arjona a,* , MªPaz Linares a , Josep Casanovas-Garcia a,b a Universitat Politècnica de Catalunya, Barcelona 08034, Spain b Barcelona Supercomputing Center, Barcelona 08034, Spain Abstract Nowadays, urban traffic affects the quality of life in cities as the problem becomes ever more exacerbated by parking issues: congestion increases due to drivers searching slots to park. An Internet of Things approach permits drivers to know the parking availability in real time. This research focuses on studying the data generated by parking systems to develop predictive models that generate forecasted information. This can be useful in improving the management of parking areas, while having an important effect on traffic. This work begins by describing the state-of-the-art parking predictive models. Then, introduces the recurrent neural network methods that were used, Long Short-Term Memory and Gated Recurrent Unit, as well as the models developed according to real scenarios in Wattens and Los Angeles. To improve the quality of the models, exogenous variables related with weather and calendar are considered. Finally, the results are described, followed by suggestions for future research. Keywords: parking availability forecast; deep learning; smart cities, recurrent models, time series. 1. Introduction Currently, one of the most important problems in urban areas concerns traffic congestion. This, in turn, has an impact on the economy, nature, human health, cities architecture, and many other facets of life. Part of the vehicular traffic in cities is caused by parking space availability. The drivers of private vehicles usually want to leave their vehicle as close as possible to their destination. However, the parking slots are limited and may not be enough to sustain the demand, especially when the destination pertains to an attractive area. Thus, individuals looking for a place to park their vehicle contribute to increasing traffic flow density on roads where the parking demand cannot be satisfied. This motivates the use of real-time parking availability in order to build a mathematical model that helps improve parking management. Some approaches to this can be found in (Caicedo et al., 2012) and in (Teodorovic and Lucic, 2006), who use models to manage indoor parking reservations while taking into account future parking demand. However, the nature of the data remains a problem: It is not on-street parking data and is not in real time. An Internet of Things (IoT) approach allows us to know the state of the parking system (availability of the parking slots) in real time through wireless networks of sensor devices. An intelligent treatment of these data could generate forecasted information that may be useful in improving management of on-street parking, thus having a notable effect on urban traffic. Smart parking systems first appeared in 2015, with platforms in Santander, San Francisco and Melbourne as is explained in (Lin et al., 2017), when those cities began to provide on-street real-time parking data for offering new services to their citizens. One of the most interesting services is parking availability forecasting, for which the first works studied the temporal and spatial correlations of parking occupancy. For this purpose, a VARIMA model was proposed by Rajabioun and Ioannou (2015) for short-term forecasts (less than 30 minutes) without loss of accuracy. * Corresponding author. Tel.: +34934011019 E-mail address: jamie.arjona@upc.edu This is an Accepted Manuscript of an article published by Taylor & Francis Group in Transportation letters: the international journal of transportation research on 25 Dec 2020, available online at: http://www.tandfonline.com/10.1080/19427867.2020.1866331