降相关对模糊度解算中搜索效率的影响分析

Impact of Decorrelation on Search Efficiency of Ambiguity Resolution

首先理论分析了条件数、正交缺陷度、S(A)等降相关评价指标所表示的几何意义,然后采用LAMBDA算法、LLL规约算法和Seysen规约算法通过模拟和实际数据对模糊度的搜索效果和不同评价指标之间的关系进行了深入计算分析。进一步验证得出"降低模糊度方差分量间的相关性实现最大程度地压缩椭球可以提高搜索效率"的观点是片面的,并通过结果分析表明提高搜索效率的本质在于尽可能地促使基向量按照一定方向排序。

The decorrelation performance of LAMBDA algorithm,LLL algorithm and Seysen algorithm are analyzed with evaluation indexes, i.e., condition number, orthogonal defect and S (A).Moreover, relationships between decorrelation performance of the above algorithms and ambiguity search efficiency are evaluated using theoretical and practical val idation,respectively.The results val idate that there is no inevitable relation between decorrelation performance of variance-covariance matrix of original ambiguity and search efficiency,whereas,traditional views consider that search efficiency can be enhanced just by improving decorrelation performance.Further analysis shows that the essence to improving search efficiency major depends on the permutation of basis vectors according to a certain di rection.