针对现存盲直接法码辅助技术抑制直接序列码分多址系统窄带干扰性能不佳的问题,提出盲子空间法码辅助技术及其自适应算法,实现对三类窄带干扰的抑制。对于音频干扰和数字窄带干扰,提出盲自适应带收缩的投影近似子空间跟踪(PASTd,Project...针对现存盲直接法码辅助技术抑制直接序列码分多址系统窄带干扰性能不佳的问题,提出盲子空间法码辅助技术及其自适应算法,实现对三类窄带干扰的抑制。对于音频干扰和数字窄带干扰,提出盲自适应带收缩的投影近似子空间跟踪(PASTd,Projection approximation subspace tracking with deflation)算法;对于AR随机过程,由于上述盲自适应算法的低秩判定困难,提出改进的盲自适应递归最小二乘(RLS,Recursive least square)预测-PASTd码辅助算法。仿真分析试验结果表明:该算法具有优越性。展开更多
The maximum of g2-d2 for linear [n, k, d; q] codes C is studied. Here d2 is the smallest size of the support of 2-dimensional subcodes of C and g2 is the smallest size of the support of 2-dimensional subcodes of C whi...The maximum of g2-d2 for linear [n, k, d; q] codes C is studied. Here d2 is the smallest size of the support of 2-dimensional subcodes of C and g2 is the smallest size of the support of 2-dimensional subcodes of C which contains a codeword of weight d. The extra cost to the greedy adversary to get two symbols of information using some algorithm is g2-d2. For codes satisfying the fullrank condition of general dimensions, upper bounds on the maximum of g2-d2 are given. Under some condition we have got code C where g2-d2 reaches the upper bound.展开更多
文摘针对现存盲直接法码辅助技术抑制直接序列码分多址系统窄带干扰性能不佳的问题,提出盲子空间法码辅助技术及其自适应算法,实现对三类窄带干扰的抑制。对于音频干扰和数字窄带干扰,提出盲自适应带收缩的投影近似子空间跟踪(PASTd,Projection approximation subspace tracking with deflation)算法;对于AR随机过程,由于上述盲自适应算法的低秩判定困难,提出改进的盲自适应递归最小二乘(RLS,Recursive least square)预测-PASTd码辅助算法。仿真分析试验结果表明:该算法具有优越性。
基金This paper was presented at International Congress of Mathematicians,August 20-28,2002,BeijingThis work was supported by the Norwegian Research Council and the National NaturalScience Foundation of China(GrantNo.10271116).
文摘The maximum of g2-d2 for linear [n, k, d; q] codes C is studied. Here d2 is the smallest size of the support of 2-dimensional subcodes of C and g2 is the smallest size of the support of 2-dimensional subcodes of C which contains a codeword of weight d. The extra cost to the greedy adversary to get two symbols of information using some algorithm is g2-d2. For codes satisfying the fullrank condition of general dimensions, upper bounds on the maximum of g2-d2 are given. Under some condition we have got code C where g2-d2 reaches the upper bound.