摘要
针对基于排序类自适应门限算法运算量大的问题,提出了利用可变的第k小元素替代排序类噪声能量估计的快速自适应门限检测算法.其中对第k小元素的取值仅采用快速排序算法的少次迭代过程,不须要排序完整的检测统计量,可实现快速噪声能量估计,以此作为自适应门限的参变量,结合相应的可变阈值因子达到快速自适应门限判决检测的目的.分析了检测概率和虚警概率以及运算复杂度,仿真对比了该算法与排序类自适应门限算法检测概率和虚警概率.结果表明:当有信号发送时,该算法与排序类检测算法检测概率基本一致,虚警概率有所减小;当没有信号发送时,相同阈值因子情况下,随着k值的增大,虚警概率越来越小.
Aiming at the problem that the caculation cost of adaptive threshold detection algorithms based on sorting is high ,and the real‐time indicator of detecting bursting signal cannot be satisfied ,a quick adaptive threshold detection algorithm was proposed ,in which the sorting estimation of noise energy was replaced by the variable k th smallest value .Instead of sorting the complete sequencing test statistic ,the propoccd quick‐sort method with few iterations was adopted to get the k th smallest value ,w hich con decreases calculation cost and realize the quick estimation of noise energy .Setting the k th smallest value as a parameter and being combined with the variable threshold coefficient ac‐cordingly could detect the signal quickly with the adaptive threshold .Detection probability ,false a‐larm probability and algorithm complexity were analyzed in theory and then the last two indexes above were compared with corresponding ones in the adaptive threshold detection algorithm with sorting method .The simulation results reveal that detection probability is in basically parallel with that of the detection algorithm and false alarm probability reduces in the presence of signal ,while false alarm probability reduces as the k enlarges in the absence of signal with the same threshold coefficient .
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2015年第8期62-65,122,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
通信网信息传输与分发技术重点实验室开放课题资助(ITD-U12003/K1260009)
航空基金资助项目(20112096016)
关键词
突发通信
自适应门限
快速排序
检测概率
虚警概率
检测算法
burst communication
adaptive threshold
quick-sort
detection probability
false alarm probability
detection algorithms
