主講人：Weihua Zhuang 院士

講座時間：5月26日 10:00-11:00

講座地點：主樓b 1421

邀請人：黃霽崴 教授

內(nèi)容簡介：Artificial intelligence (Al) models will continue to be pervasively deployed to supportdiverse applications in the 5G/6G era. In this presentation, we investigate Al for cooperative perceptionto achieve reliable situation awareness of connected and autonomous vehicles (CAVs). We discuss thekey challenges and present an accuracy-aware and resource-efficient raw-level cooperative sensing andcomputing scheme among CAVs and road-side infrastructure. A supervised learning model is trained tocapture the relationship between the object classification accuracy and the data quality of selected objectsensing data, facilitating accuracy-aware sensing data selection. We formulate an optimization problemfor joint sensing data selection, object classification task placement, and resource allocation, to minimizethe total resource cost while satisfying the delay and accuracy requirements. A genetic algorithm basediterative solution is proposed for the optimization problem, Numerical results demonstrate the accuracyawareness and resource efficiency achieved by the proposed cooperative sensing and computing scheme.in comparison with benchmark solutions.