Parallel Computing for Processing Data from Intelligent Transportation Systems

Abstract

This article describes the application of parallel computing techniques for efficiently processing large volumes of data from ITS. This is a relevant problem in nowadays societies, especially when working under the novel paradigm of smart cities. The proposed approach applies parallel multithreading computing for processing Global Positioning System records for a case study on the Intelligent Transportation System in Montevideo, Uruguay. The experimental analysis is performed on a high performance computing platform, considering a large volume of data and different computing resources. The main results indicate that the proposed approach allows achieving good speedup values, thus reducing the execution time to process more than 120 GB of data from 921 to 77 min, when using 32 threads. In addition, a web application to illustrate the results of the proposed approach for computing the average speed of public transportation in Montevideo, Uruguay, is described.