Impact of Decorrelation on Search Efficiency of Ambiguity Resolution
LU Liguo, Wanke Liu, Jiangwei Li · DOAJ (DOAJ: Directory of Open Access Journals) · 2015
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 validation, respectively. The results validate 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 direction.