UAV-Assisted Vehicle Network Task Offloading System for Multi-Traffic Scenarios

With the growth of the low-altitude digital economy, Unmanned Aerial Vehicle (UAV) has been introduced into Intelligent Traffic System (ITS) to build UAV-assisted vehicular networking networks. The UAV-assisted Mobile Edge Computing (MEC) System, which combines the flexibility of UAVs with the low-latency advantages of edge computing, has garnered widespread attention in recent years. However, the UAV’s computing resources are limited and cannot handle large data tasks. In areas lacking communication infrastructure, task processing requires queuing, which fails to meet the real-time requirements of the tasks. In order to address the issue, this paper proposes a task offloading strategy proposes a task offloading strategy for collaborative offloading between UAV and Roadside Computing Unit (RCU). The Deep Deterministic Policy Gradient (DDPG) algorithm is employed for iterative optimization, ultimately yielding offloading coefficients that minimize the weighted sum of normalized delay and energy consumption, thereby deriving a more efficient task offloading strategy with low delay and energy consumption.