Mathematical Modeling and Simulation of Traffic Flow Control in Urban Environments Using Fuzzy Deep Neural Network With Optimization Algorithm

With the enlarging number of transports on the road and fast growth, traffic flow is a significant
current worry that obstructs the financial system’s evolution and affects the quality of life. Intelligent
Transportation Systems (ITS) utilize innovative technologies in urban environments to enhance urban and
interurban traffic, decrease congestion, and improve overall traffic flow control. A usual method employed
for traffic flow control is usually based upon the analysis and collection of data in a physical manner
that is energy-demanding and tedious. Currently, with the improvements in Machine Learning (ML), Deep
Learning (DL), and Artificial Intelligence (AI), urban environments are observed to induce concerns of the
environment properly, with the optimum control of traffic pollution, congestion, and other effects. Hence, this
article designs a Traffic Flow Control utilizing a Fuzzy Deep Neural Network with a crayfish optimization
algorithm (TFCFDNN-COA) technique in Urban Environments. The TFCFDNN-COA technique mainly
aims to control traffic flow levels in urban environments, allowing efficient traffic management. At first, the
TFCFDNN-COA approach includes data pre-processing and a dingo optimizer algorithm (DOA) model for
feature selection. The fuzzy deep neural network (FDNN) technique controls traffic flow. Eventually, the
crayfish optimization algorithm (COA) model is utilized to fine-tune the best hyperparameter of the FDNN
model. A wide range of experimental studies has been completed, and the outcomes have been studied using
numerous features. The experimental validation of the TFCFDNN-COA approach portrayed superior MSE,
MAE, and MAPE values of 0.0011, 0.0206, and 0.6738, respectively, all observed on Sunday.