多用户OFDM系统基于压缩感知的信道估计中多导频集合设计研究

Design of multiple pilot sets for compressed sensing based channel estimation in multi-user OFDM systems

为了保证高速移动环境中多用户OFDM系统(Multi-User Orthogonal Frequency Division Multiplexing Systems)中基站能估计所有的上行信道,研究了双选择(Doubly-Selective,DS)信道模型下多个导频集合的设计问题。同时为了降低每个用户的导频开销,将压缩感知(Compressed Sensing,CS)理论应用在信道估计中。在分析基于最小化互相关值的导频设计准则所存在问题的基础上,将最小化调整互相关值用于传统的DSO(Discrete Stochastic Optimization)导频设计方法来完成单个用户导频设计,然后基于最小化调整互相关集合元素的加权平均准则,将改进的DSO算法扩展到多用户导频集合的设计。仿真结果表明,采用提出的新准则获得的每一个导频集合均能获得比较好的信道估计性能,与已有的其它准则相比,在用户数增多的情况下能使信道估计性能得到改善。

To ensure that base stations in multi-user orthogonal frequency division multiplexing （OFDM） systems can identify all channels in high-speed mobile environments, the design of pilots for multiple users in doubly-selective （DS） channel estimation is discussed. To reduce the pilot overhead of every user, compressed sensing （CS） theory is applied to channel estimation. Since the pilot design criterion based on minimizing the cross-correlation value is deficient, the pilot design criterion is modified in the discrete sto- chastic optimization （DSO） algorithm to complete the pilot design for single user. By using the criterion of minimizing weighted average to modify the cross-correlation value, the improved DSO algorithm is extend- ed to the pilot design for muhi-users. Simulation results show that each pilot set obtained by the scheme can obtain superior channel estimation performance, and a smaller mean square error （MSE） than other current optimization algorithms when the number of users increases.