Traffic Prediction in Smart Cities Based on Hybrid Feature Space

In smart cities of the future, data will be generated, integrated, processed and utilized from heterogeneous sources at varying levels complexity. For urban traffic planning in cities, one biggest challenges is congestion prediction its avoidance. Traffic a complex phenomenon it manifestation various contributing factors. addition to vehicular mobility, properties road network, weather, holidays peak hours play significant role especially on arterial roads within city. this paper, we proposed hybrid GRU-LSTM based deep learning model applied city-wide novel integrated sources. We have devised our indigenous pipeline that composed set algorithms dealing with map matching, sparsity handling, outlier removal, zero speed adjustments, Open Street Map (OSM) segment mapping etc. Extensive experimentations been carried out demonstrate improved performance method. The comparative analysis reveals methodology yields 95 % accuracy outperforms other neural network models.