Cycled efficient V-BLAST GNSS ambiguity decorrelation and search complexity estimation

Abstract

Decorrelation reduction and search techniques play important roles in solving integer ambiguities of global navigation satellite system (GNSS) positioning. This contribution focuses on these two issues and can mainly be divided into two parts. (1) An efficient strategy to realize the popular vertical Bell Labs layered space–time (V-BLAST) algorithm is studied. Then, a cycled efficient V-BLAST (ceV-BLAST) reduction algorithm is constructed. The decorrelation outcome of ceV-BLAST reduction is demonstrated by analysis to be stronger than LAMBDA and LLL. (2) A search complexity estimator which can give a priori estimation of the search complexity is proposed. This estimator evaluates the search nod number directly and turns out to follow the real search complexity better than other estimators of the same type. At the same time, it is very time efficient. Utilizing two sets of GNSS observation data, we compare ceV-BLAST and the new search complexity estimator with existing decorrelation algorithms and estimators, respectively. The results have confirmed our arguments.