RSMA使能的通感算一体化系统公平性计算卸载优化
Fair computation offloading optimization for rate‑splitting multiple access enabled integrated sensing, communication,and computing systems
  
DOI:
中文关键词:  波束赋形设计;通感算一体化;多层计算;速率拆分多址接入
英文关键词:beamforming design; integrated sensing, communication, and computing (ISCC); multitier computing; rate splitting multiple access (RSMA)
基金项目:国家自然科学基金(62071255)和江苏省重点研发计划(BE2023087)资助项目
作者单位
齐婷 南京邮电大学 宽带无线通信与传感网技术教育部重点实验室,江苏 南京 210003 
顾雨航 南京邮电大学 宽带无线通信与传感网技术教育部重点实验室,江苏 南京 210003 
吴俊豪 南京邮电大学 宽带无线通信与传感网技术教育部重点实验室,江苏 南京 210003 
王磊 南京邮电大学 宽带无线通信与传感网技术教育部重点实验室,江苏 南京 210003 
摘要点击次数: 156
全文下载次数: 222
中文摘要:
      针对通感算一体化(Integrated Sensing, Communication, and Computing, ISCC)系统中通信、感 知与计算三大功能高效协同的问题,文中提出一种基于速率拆分多址接入(Rate Splitting Multiple Access, RSMA)的 ISCC 系统框架。在该框架中,多功能全双工基站(Multifunctional Full-Duplex Base Station, MFDBS)引入 RSMA 机制,不仅实现了高精度目标感知,还能协同处理邻近用户的边 缘计算任务。相较传统多址技术,RSMA通过灵活划分公共流与私有流的功率资源,展现出更强的 干扰管理能力与传输鲁棒性,特别是在非完美串行干扰消除(Successive Interference Cancellation, SIC)条件下仍保持优异的性能稳定性。在满足通信-计算因果性与感知质量约束的前提下,以最 大化最小用户计算速率为目标,构建联合优化模型,协同设计MFDBS的发送波束赋形与RSMA功 率分配策略。针对该非凸优化问题,进一步提出一种基于加权最小均方误差的交替优化算法进行 求解。仿真结果表明,所提 RSMA优化算法在非完美 SIC条件下显著优于非正交多址接入与空分 多址接入方案,其灵活的公共-私有流分层机制可协调感知干扰与计算资源竞争,使最小用户计算 速率在MFDBS功率、最小感知信噪比及用户数量变化时均有提升。算法收敛性分析进一步揭示, RSMA通过动态功率分配与干扰主动抑制,在保障感知性能的同时突破传统方案的上行速率瓶颈, 展现出RSMA在通感算一体化系统中的强鲁棒性。
英文摘要:
      To address the challenge of efficiently coordinating communication, sensing, and computing in integrated sensing, communication, and computing (ISCC) systems, a rate splitting multiple access(RSMA)-based ISCC framework is proposed. In this framework, a multifunctional full-duplex base station (MFDBS) is designed with RSMA, enabling simultaneous high-precision target sensing and collaborative processing of edge computing tasks for nearby users. Compared with conventional multiple access schemes, RSMA demonstrats superior interference management and transmission robustness through flexible power allocation between common and private streams, particularly under imperfect successive interference cancellation (SIC) conditions. A joint optimization model is formulated to maximize the minimum user computation rate under communication-computation causality and sensing quality constraints, where the transmit beamforming of MFDBS and RSMA power allocation are co-designed. To solve the non-convex optimization problem, a weighted minimum mean square error (WMMSE)-based alternating optimization algorithm is developed. Simulation results validate that the proposed RSMA algorithm significantly outperforms non-orthogonal multiple access (NOMA) and space division multiple access (SDMA) schemes under imperfect SIC. The hierarchical common-private stream mechanism is shown to effectively mitigate sensing interference and resolve computing resource competition, achieving consistent improvements in the minimum user computation rate across varying MFDBS transmit power, minimum sensing signal-to-noise ratio, and user numbers. Convergence analysis further reveals that RSMA breaks the uplink rate bottleneck of traditional schemes via dynamic power allocation and proactive interference suppression while guaranteeing sensing performance, demonstrating its strong robustness in ISCC systems.
查看全文  查看/发表评论   附件

你是第5855880访问者
版权所有《南京邮电大学学报(自然科学版)》编辑部
Tel:86-25-85866913 E-mail:xb@njupt.edu.cn
技术支持:本系统由北京勤云科技发展有限公司设计