Real-Time AI-Based Traffic Prediction for Autonomous Vehicles

Abstract

As part of the core effort to make transportation safe, effective, and efficient, mild or fully self-driving vehicles are being developed and tested around the world. Autonomous transport systems in cities can potentially benefit from other technical areas, including integrated traffic management, greatly boosting their effectiveness and, by doing so, the quality of service and safety. AI is becoming one of the pillars of upcoming traffic systems. AI can be used to build predictive models of traffic, which are useful for the management strategies to maintain traffic within operational boundaries. The more timely such predictions are, the better traffic systems can adapt. It has, thus, become increasingly clear that real-time traffic prediction that can work effectively will be fundamental in a future world with ever-increasing levels of automation. For urban intelligent transport systems whose designs increasingly rely on AI-based methods, traffic analysis is one of—if not the—primary data sources used.