Hierarchical Edge AI for Cooperative Driving

This paper presents vehicle-to-vehicle (V2V) co-operation with hierarchical edge AI and we demonstrate it with real-vehicle experiments. We propose a Wi-Fi and MQTT communication framework, implement it, and evaluate it through platooning experiments on public roads. We confirm that V2V co-operative processing meets real-time requirements while reducing communication to tens of kbps with sub-10 ms latency. For the demo, we show it with real-vehicle experiments, highlighting the novelty of our approach, its real-time capability, and its potential to engage the audience.