Neuromorphic computing architectures for enhanced signal processing in autonomous systems

Abstract

The increasing complexity of urban traffic management demands innovative solutions to enhance traffic flow, minimize congestion, and improve safety. Traditional methods often fall short in handling real-time traffic prediction and signal optimization. This study proposes a hybrid approach that combines Convolutional Neural Networks (CNNs) with the Internet of Things (IoT) framework to predict traffic flow and optimize traffic signals dynamically in smart cities. Data from IoT sensors installed across urban locations is fed into a CNN model, which forecasts traffic patterns and adjusts signal timings accordingly. Our findings show that the CNN-IoT model achieves 92.5% accuracy in traffic flow prediction, reduces congestion by 18%, and enhances signal efficiency by 15%. The system is capable of adapting to unforeseen traffic conditions, demonstrating significant improvements over traditional methods. This study highlights the potential of combining deep learning and IoT for real-time traffic management and sets the stage for scaling such systems in larger urban settings for further validation