The identification of road modality and occupancy patterns by Wi-Fi monitoring sensors as a way to support the “Smart Cities” concept

Abstract

As world urbanization continues apace and total population increases, there is an immediate demand for better monitoring and exploitation of space. In view of the above, the “Smart Cities” concept has been developed and numerous efforts are made to deploy technology to this end. The main information needed for city development and planning is road modality and the relevant occupancy patterns. However, it is quite difficult to collect this information. There have been several different approaches towards providing this information and various methods have been used. However, each of them has weaknesses which do not allow it to be used on its own. On the other hand, thanks to the new technological developments and due to the growing needs of society over the last years, the system of Wi-Fi monitoring sensors has been increasingly used in outdoor environments. Many companies have already used this method to collect data and provide information about users’ behavior in places such as public areas, shopping centers and malls. Nonetheless, the contribution of this thesis is the study of the applicability of this method, the assessment of the reliability of its outcomes and the identification of crucial parameters which significantly affect the final accuracy. Thus, the aim of this research is to investigate what kind of road modality and occupancy patterns can be recognized using Wi-Fi monitoring sensors in a city area as well as which setup parameters can influence the final outcome. The system is implemented in the city of Dordrecht, which constitutes the research area of this study. First of all, the design of the observation network is described and the relevant parameters are taken into account. Using the data collected by the system and the known distances between the sensors, the movement speed of each device is computed. Street-uses criteria of the research area are also used as input to the system, and in combination with the computed speed three categories of users are recognized and each device is categorized as “pedestrian”, “bicyclist”, or “vehicle”. Under this classification each street’s road modality is studied. The relationship between the categories throughout the day is investigated and preferred streets for each kind of users are recognized. Based on the ability of the system to identify every device in the research area throughout the day, the movement behaviors of users are researched and similarities between them as well and the most frequent patterns are identified. Three sets of movement patterns are studied considering the number of sensors which scan the same device within a time period. Each set is investigated separately for every kind of users. Moreover, using the number of devices scanned at each sensor point, occupancy patterns are identified both for users as a whole and for each user category separately. It is argued that this constitutes an important advantage of the system. Rush hours, recession periods and movement trends are recognized for the different days of the week as well as the occupancy relationship between the research area and its surroundings. Finally, a questionnaire and random samplings with Bernoulli trial are used to validate the outcomes. A quite strong correlation between the system’s results and reality is revealed, especially with regard to pedestrians and bicyclists. However, despite the quite promising findings, further implementation and testing of the system in different environments is needed in order to draw an indisputable conclusion about its effectiveness.