面向大规模接入的通信感知一体化

Integrated sensing and communication for massive access

随着万物互联的愿景不断推进,无线电频谱稀缺已成为制约未来无线通信系统发展的严峻挑战.为了解决上述问题,通信感知一体化(integrated sensing and communication,ISAC)逐渐成为第六代(sixth-generation,6G)移动通信技术与业务的主导趋势之一.ISAC系统不仅能完成可靠的多设备通信,同时能实现准确的感知,有望成倍提高频谱效率.同时,具有高可靠、低时延和小成本特性的大规模随机接入技术被广泛认为是6G通信网络的核心技术之一.为此,本文提出了面向大规模随机接入的通信感知一体化框架,即大规模随机接入系统可以利用相同的频谱、硬件和信号处理模块来完成物体成像的感知服务.具体而言,激活设备同时向基站发送待解码的数据信息,信号被环境中物体散射后到达基站.在基站端,本文探究了大规模接入的偶发性通信的特性和环境信息的稀疏特性,基于张量建模与分解、变分贝叶斯(Bayes)推断和字典学习技术,提出了新的联合激活设备检测和物体成像感知算法.严谨的理论分析和大量的仿真实验验证了所提算法的准确性和高效性.

With the widespread application of the Internet-of-everything,the radio spectrum crunch has become a severe problem for wireless communications.To solve this problem,integrated sensing and communication(ISAC) has gradually become one of the key trends of sixth-generation(6G) wireless communication technology.ISAC requires the system to complete reliable multi-user communication and achieve accurate sensing,which is expected to significantly improve spectral efficiency.Moreover,massive random access with high reliability,low latency,and low cost is widely regarded as one of the key technologies of 6G wireless networks.Thus,this paper proposes an ISAC framework for massive random access,in which massive random access systems can utilize the same spectrum,hardware,and signal processing modules to provide object sensing services.In particular,active devices send unknown data to a base station(BS),and the data signal hits the object and results in signals bouncing off of objects before they arrive at the BS.This paper explores the characteristics of sporadic communication in massive access and sparsity in environmental information.Based on tensor modeling and decomposition,variational Bayesian inference,and dictionary learning techniques,an algorithm is proposed for object sensing and the detection of novel joint device activity.The accuracy and efficiency of the proposed active device detecting and imaging algorithm were verified by rigorous theoretical analysis and a large number of simulation results.