Search

For the requirements of real-time services in autonomous driving systems, such as high-definition (HD) maps, based on the integration of sensing, communication, and computing, a multi-node cooperative sensing mechanism was proposed with the age of information (AoI) as the real-time indicator.Considering the constraints on communication and computing resources and vehicle energy consumption, the information collection, transmission and processing of sensing nodes were optimally scheduled to minimize the AoI averaged over time.A Lyapunov-based online scheduling algorithm was proposed to transform the long-term stochastic optimization problem into an online optimization problem, which could be solved with low complexity.Compared with the existing mechanism considering integrated communication and computing, the proposed mechanism can improve real-time performance by 9%～50%.

Massive multiple-input multiple-output (mMIMO) enables higher spectrum efficiency by deploying a large number of antenna elements.However, mMIMO will bring a substantial increase in the number of radio frequency (RF) channel, which will result in a sharp increase in cost.Therefore, it is necessary to design mMIMO with comprehensive consideration of the trade-off between the performance and the cost.Sparse array technology can approach the performance of the uniform array with less number of antenna elements by optimizing the position distribution and excitation amplitudes of the antenna elements, thereby reducing the number of RF channels and controlling the cost, and achieving a better cost-effective balance.Both link-level and system-level evaluations and analysis of sparse array based mMIMO that targets for 6G were performed under realistic simulation settings.The results show that with little performance loss, the sparse array based mMIMO can significantly reduce the number of antenna elements and the cost of base station infrastructure.

6G will evolve to an intelligent network, featured by native intelligence and openness. The study of standardization of intelligent network emphasizes the importance of intent-driven network for network intelligence. However, current intent-based networking takes the intent as "What to do" rather than "What you want". Moreover, current knowledge defined network (KDN) can partially fulfill "How to configure the network" according to "What to do". Therefore, a 6G native intelligent network architecture based on the intent abstraction and knowledge was proposed, aiming to achieve "How to configure the network" according to "What you want". Firstly, an intent abstraction module was designed to obtain "What to do" from "What you want", composed of intent acquisition, intent translation, intent mapping, and intent modeling. Secondly, the cognitive module was proposed to achieve "How to configure the network" according to "What to do", which got network knowledge through joint optimization of machine learning and logical reasoning. Finally, enabling key technologies such as intent mapping, network information measurement, network policy generation, and network policy verification were introduced to support the implementation of 6G native intelligence. [ABSTRACT FROM AUTHOR]