一种粒子滤波的盲多用户检测快速算法  

A Fast Algorithm of Blind Multiuser Detection Using Particle Filtering

作　　者：田红波[1] 殷勤业[1] 丁乐[1] 邓科[1] 

机构地区：[1]西安交通大学电子与信息工程学院

出　　处：《电子与信息学报》2008年第6期1300-1303,共4页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60572046,60502022);教育部博士点基金(20030698027)资助课题

摘　　要：该文在时间观察状态空间模型基础上,针对粒子滤波计算复杂的问题,构造了一种在同步快变平衰落信道下计算复杂度低的粒子滤波的DS-CDMA盲多用户检测快速算法。该算法通过对超过门限数目的粒子分类,以不同类别粒子的概率差值作为粒子滤波所需粒子数目是否足够的判断准则,自动调节不同情况下所需的粒子数目,并对不同类别粒子概率差值下盲多用户检测性能进行分析。仿真结果表明,该文算法可以在保证盲多用户检测性能的前提下有效地降低运算复杂度,且检测性能与概率差值成正比。Considering the computational complexity of particle filtering, based on the time-observation state space model, a new fast algorithm with low computational complexity is developed for DS-CDMA blind multiuser detection using particle filtering in synchronous systems over flat and fast fading channels. This algorithm classifies the particles when its number exceeds the threshold, and the probability difference of different particles is taken as the criterion whether the particles number is enough, and then the particles number under different circumstance is adjusted adaptively; meanwhile the performance of blind multiuser detection under distinct probability difference of different particles is discussed. Simulation results confirm that this algorithm effectively decreases computational complexity and well remains the performance of blind multiuser detection, meanwhile the performance in proportion to the probability difference.

关 键 词：盲多用户检测 计算复杂度 粒子滤波 快速算法 

分 类 号：TN911.7[电子电信—通信与信息系统]