A Real-Time Traffic Management Framework for Smart Cities Based on Fuzzy Neural Networks

Smart traffic management is essential for minimizing congestion, increasing road safety, andimproving overall transportation efficiency in modern smart cities. In this study, five trafficmanagement systems Adaptive Traffic Signal Control, Intelligent Transportation System,Automatic Number Plate Recognition, Smart Pedestrian Crossing, and Emergency VehiclePriority System are evaluated using a back propagation-based interval valued Fermateanpicture fuzzy (IVFPF) neural network model.Primary performance indicators, including response time, accuracy, energy efficiency,reduction in traffic delays, and cost-effectiveness, are treated as macro-level factors.Supporting elements such as advanced traffic control signals, pedestrian sensors, smartsignaling infrastructure, pedestrian movement detectors, and intelligent monitoring systemsare considered micro level factors.At the initial stage, system performance was assessed using only macro level parameters;however, the outcomes did not achieve the expected efficiency level. By graduallyincorporating micro-level factors and employing back propagation for iterative errorminimization, the systems performance improved step by step. After multiple trainingepochs, an optimal balance between macro and micro parameters was achieved, resulting inmaximum overall efficiency of the traffic management system.Additionally, Hebbian learning is explored as a mechanism in which neural connections aremodified purely based on the correlation of neuron activations, allowing synaptic strengths toincrease or decrease without the use of explicit error feedback.Among the five traffic management systems studied, the Emergency Vehicle PrioritySystem is found to be more reliable and effective than the others.