基于概率排序的存储约束树形搜索算法的研究*1

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。

Research of Memory Constrained Tree Search based on Probabilistic Sorting

Considering the currentSMIMOSsystem shortcomings of largeSstorage space requirementsSand highScomplexity inSmultiple-symbol differentialSdetection algorithm,Sthis paper proposes aSprobabilisticSsortingSmemorySconstrained tree search algorithmS(PSMCTS),Susing performanceSadvantage of sorting algorithm and storageSadvantage of MCTS, to solve this problem. ThroughStheoretical analysis and simulation,SPSMCTS canSeffectively inherit the MCTSSalgorithm goodSadvantage,Sdynamically adapt toSthe preset storageSspace,Sand suitable for hardware implementation.SUsing sorting algorithmsSimproved the detectionSperformance,Sand the performanceSis close toSML algorithm under fixed memory requirement. The algorithm also solves the high computational complexitySproblem ofSMCTSSalgorithm inSsmall storage capacitySconditionsSunder the lowSSNR region. Therefore,SPSMCTSSisSa goodSschemeSin communication systems.