Classification of Traffic Pattern and Estimation of Traffic Speed Using Recurrent Neural Network in Smart Cities

It is important to accurately manage traffic jams in metropolitan cities, which can be challenging because of complicated and changing traffic trends. An intelligent transport structure (ITS) integrates novel detecting, management, and communication innovations to increase traffic performance. To monitor the dynamic status of traffic size, a crucial generated variable that is required is traffic density forecasting (TDF). Increasing numbers of automobiles with wireless connections open up novel possibilities for current and more precise TDF than earlier approaches. This study investigates the assessment of traffic density employing a limited set of basic indicators, like the number of automobiles in the vicinity and the distribution of beacons from roadway units to adjacent units and inversely. It is designed to be used for the analysis of actual traffic tracks, variable estimation for models, and the creation of traffic simulations. This study incorporates a spatial-temporal GGAN algorithm, for real-time traffic density forecasting utilizing the PMS4 and PMS8 real-traffic databases. The empirical findings support the 96.6% and 98.7% accuracy scores for the PMS4 and PMS8 databases, accordingly.