Maximum Likelihood based Direct Position Estimation for Mobile Stations in Dense Multipath
share
The problem of direct position estimation in dense-multipath mobile scenarios is addressed. A low-complexity, fully-adaptive algorithm is proposed, based on the maximum likelihood approach. The processing is done exclusively on-board at the mobile node by exploiting the training sequence of a narrowband radio signal; thus, an antenna array is required only on the mobile node, as opposed to MIMO approaches where multiple antennas are considered at both sides. The proposed algorithm is able to estimate via adaptive beamforming (with spatial smoothing for coherence decorrelation) the optimal projection matrix that makes the received signal orthogonal to the multipath subspace; in addition, it tracks line-of-sight associations over the trajectory, hence achieving an integration gain. A simpler variant of the algorithm is also devised, which does not require any training sequence, at the price of some accuracy loss. The performance assessment shows that the proposed algorithms are very effective in (dense) multipath conditions, significantly outperforming natural competitors also when the number of antennas and snapshots is kept at the theoretical minimum.
READ FULL TEXT