Intelligence-Sharing Vehicular Networks with Mobile Edge Computing and Spatiotemporal Knowledge Transfer
Jie Guo, Wenwen Luo, Bin Song, F. Richard Yu, Xiaojiang Du
Abstract
Based on recent advances in MEC and knowledge transfer in artificial intelligence, we propose a novel framework named ISVN, in which the intelligence of different MEC servers can be shared to improve performance. Specifically, we present the main techniques in the ISVN framework, including aggregation and representation for context features, relationship mining and reasoning, and knowledge transfer among MEC servers. The results of object detection experiments with the proposed ISVN framework are presented. By taking advantage of MEC and knowledge transfer, the processing speed and accuracy of object detection can be significantly improved in different scenarios of vehicular networks.