Autonomous Driving Optimization through Cognitive IoT in Intelligent Transportation Systems

Abstract

This study proposes a multi-layered Cognitive Internet of Things (CIoT) framework to optimize autonomous driving within Intelligent Transportation Systems (ITS). The framework integrates IoT sensing, real-time edge computing, cloud-based analytics, and artificial intelligence to enhance traffic flow, safety, and sustainability. The architecture consists of five collaborative layers—IoT, data, cognitive computing, cloud computing, and service—each designed to process, analyze, and respond to dynamic traffic conditions. Machine learning techniques such as LSTM, CNN, and reinforcement learning are deployed for adaptive signal control and congestion prediction. Simulation experiments conducted using VISSIM and SUMO show significant improvements in vehicle delay, throughput, travel time, fuel efficiency, and emissions. The proposed system provides a scalable and secure foundation for next-generation smart mobility, effectively supporting autonomous and connected vehicle ecosystems.