Polling Resource Scheduling for Internet of Vehicles Based on U-Net Semantic Segmentation

With the continuous development of Internet of vehicles (IoV) and intelligent transportation system (ITS), infor-mation interaction and resource sharing between vehicles have become increasingly important. However, the resource scheduling problem in IoV has been the main bottleneck restricting its fur-ther development. This paper proposes a semantic segmentation method based on U-Net deep learning architecture and applies it to the polling scheduling problem in the IoV environment to achieve the optimal allocation of resources. The simulation results show that compared with the traditional polling scheduling method, the semantic segmentation method combined with U-Net can significantly save bandwidth resources and reduce image transmission delay. Specifically, our algorithm achieves a 97 % bandwidth savings and a 90% reduction in image transmission delay. In the corresponding semantic segmentation algorithm, the U - Net algorithm also outperforms the SegNet algorithm model in real-time transmission of Internet of Vehicles image resources, which verifies its potential in resource scheduling optimization.