摘要
由于传统的马尔可夫链蒙特卡罗(MCMC)方法在高信噪比或迭代过程中,容易"陷入"某一采样状态从而影响到检测性能,本文提出一种改进的MCMC方法,基于最小均方误差(MMSE)检测算法确定采样初值,使马尔可夫链迅速收敛;对陷入状态下的采样序列,随机反转具有较大后验方差的比特,以增加有效的采样状态数.仿真结果表明,改进MCMC算法能改善系统的误码率(BER)性能、降低运算复杂度.
Because the general Markov Chain Monte Carlo （MCMC） method is easily to get trapped in some fixed state especially in the case of high SNR condition or iteration process that leads t6 performance degrada- tion, in this paper, an improved MCMC algorithm is presented. Determining the initial sample value based on MMSE principle, the Markov chain is convergent rapidly. Simulation results show that the mentioned ＇algorithm can imorove the BER oerformance and reduce comnutatinnal cnmnlaxilv
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
2012年第9期96-100,共5页
Journal of Harbin Institute of Technology
基金
国家自然科学基金资助项目(61172132)
浙江省教育厅科研计划资助项目(KYF045609004)
