Statistical Properties of Vehicle Residence Times for Fog Computing Applications

Currently, new advances in the automotive industry are focused on implementing autonomous cars, since they are the future to avoid accidents and protect users. Even if safety is the main goal, many opportunities for technological developers are possible. Among them, Smart Cities is a major player in future communication networks. In a smart city context, hundreds or thousands of sensors will be deployed in many strategic parts of the city, including sensors in mobile devices, that can provide critical information for the city management and improve the resident’s livelihood. However, this scenario entails extremely high volumes of information to be sent to different geographical locations. Because of this, the use of cellular base stations may be a highly expensive alternative. In this work. We propose to take advantage of the use of autonomous cars as data mules for the efficient recollection of data in smart cities environments. Specifically, we consider interest points in Luxembourg City, where relevant data may be generated by sensors in mobile devices or fixed sensors in the city’s infrastructure. Assuming that, autonomous vehicles know in advance the route that they are going to follow to reach their destination, sensors can profit the passage of these vehicles to transmit their data, making short-range, low-cost transmissions and reducing the implementation cost of these applications. Later, the vehicles can relay the data on the destination point. To this end, we evaluate the potential use of this system by obtaining the main statistics variables of the passage of the vehicles though these interest points in the city. We obtain the mean, variance and coefficient of variation of the resident times of vehicles to estimate the potential use of this communication system in Smart Cities.