Towards theory of generic Principal Component Analysis
Journal of Multivariate Analysis
Generic weighted filtering of stochastic signals
IEEE Transactions on Signal Processing
| Hi-index | 35.69 |
In mobile communications, coherent detection requires the estimate of an increasing number of channel parameters due to the promising performance of systems that deploy multiple antennas. However, the estimate of space-time channels requires a number of training symbols that grows with the number of unknowns. To overcome this problem, we propose a subspace-based estimation method that exploits the different varying rates in the structure of the space-time channel for moving terminals. Since the channel has some fast-varying (faded amplitudes of the paths) and slowly varying (delays and directions of arrival) features, the multislot (MS) estimate is composed of two terms: the slowly varying spatial and temporal bases estimated from L consecutive slots and the fast-varying amplitudes estimated on a slot-by-slot basis. Performance analysis and simulations confirm the expected benefits of the multislot approach and demonstrate that for large L, the mean square error (MSE) on the channel estimate depends only on the number of fast-varying parameters.