基于概率排序的存储约束树形搜索算法的研究  

Research of Memory Constrained Tree Search Based on Probabilistic Sorting

作　　者：朱瑞鑫 金小萍[1] 冯会真[1] 

机构地区：[1]中国计量学院信息工程学院,浙江杭州310018

出　　处：《电视技术》2014年第23期124-127,共4页Video Engineering

基　　金：国家自然科学基金项目(61071119);浙江省自然科学基金项目(Y1091155;LQ12F01010);东南大学国家移动通信研究实验室开放性研究基金项目(2011D18)

摘　　要：鉴于目前MIMO系统中大多数多符号差分检测算法对于大容量存储空间的需求和高计算复杂度的缺点,提出了一种概率排序的存储约束树搜索(Probabilistic Sorting Memory Constrained Tree Search,PSMCTS)算法,利用概率排序的性能优势与MCTS的存储优势来解决此问题。经过理论分析与仿真验证,该算法能够继承MCTS算法的优势,能够动态地适应预设的存储空间,适合硬件实现,而排序算法提高了检测性能,在固定的存储需求下,性能表现更加逼近ML算法,同时能够解决MCTS算法在小存储容量条件下低信噪比区域计算复杂度仍比较高的问题。因此,PSMCTS可以作为一种有效的方案应用在通信系统中。Considering the current MIMO system shortcomings of large storage space requirements and high complexity in multiple-symbol differential detection algorithm, a probabilistic sorting memory constrained tree search algorithm （PSMCTS） using performance advantage of sorting algorithm and storage advantage of MCTS is proposed. Through theoretical analysis and simulation, PSMCTS can effectively inherit the MCTS algorithm good advantage, dynamically adapt to the preset storage space, and suitable for hardware implementation. Using sorting algorithms improved the detection perfotraance, and the performance is close to ML algorithm under fixed memory requirement. The algorithm also solves the high computational complexity problem of MCTS algorithm in small storage capacity conditions under the low SNR region. Therefore, PSMCTS is a good scheme in communication systems.

关 键 词：MIMO 概率排序 存储约束属性搜索 

分 类 号：TN911.23[电子电信—通信与信息系统]