An Improved DBSCAN and Multi-Agent Based Task Offloading Mechanism for 6G-Enabled Internet of Vehicles

High mobility of Internet of Vehicles (IoV) brings rapidly changing network topology, and massive data produced by vehicles aggravate heavy burden to the network. These may lead unreliable and high latency of data transmission and processing, which is not facilitate the application and popularization of automatic driving. Evolutions of intelligent vehicles and edge intelligence promising technologies enable vehicles as agents. Vehicles have abilities to act as aided Mobile Edge Computing (MEC) servers to support ultra-low communication and computing latency and super-high reliability data transmission and processing. In this paper, we elect some vehicles as aided MEC servers and design an improved Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering algorithm to involve more scattered vehicles for clustering. We adopt a Multi- Multi matching algorithm to pair vehicles and the aided MEC server, and designed a multi -agent- based task offloading mechanism to reduce latency and improve resource utilization efficiency. Furthermore, a reward mechanism is proposed to stimulate vehicles to be aided MEC servers instead of refusing to provide services. Evaluation results verify the proposed offloading method could effectively reduce the average delay, improve computing resources utilization and increase the benefit of aided MEC servers and service providers.