On the tracking performance of least squares MIMO channel estimation in rician fading

Abstract

Abstract:The tracking Mean Square Error (MSE) of the least squares (LS) Multiple Input Multiple Output (MIMO) channel estimation in rician flat fading channel is derived in closed-form, which is then verified by Monte Carlo simulation. The channel accomodates applications where the elevation angle between the transmitter and the receiver dictates the doppler shift of the Line of Sight (LOS) component, as encountered in communications from satellite or High Altitude Platform (HAP) to a moving mobile. A higher tracking loss is revealed as the mobile travels to a lower elevation angle.

Performance Improvement of LeastSquares Adaptive Filter for High-Speed Train Communication Systems

“…Zakia, et al have shown that for a moving receiver, the Doppler shift of the LOS channel component ( d 90°) yields worse MSE tracking performance at the receiver [6]. Specifically, the MSE tracking gain achieved when the mobile travels from 0° to 60° is insignificant if compared to when it travels to a 90° elevation angle.…”

“…Faletti, et al [2] have implemented the multiple-input multiple-output (MIMO) beamforming recursive least-squares (RLS) algorithm, while examples of single-input multiple-output (SIMO) are given by White, et al in [4] and Zakia, et al In [5], the recursive LS (RLS) MIMO channel estimation algorithm for Rician fading was proposed by Zakia, et al [6] as a generalization to the Rayleigh fading case given in Karami [7]. Both papers present closed-form solutions on the tracking performance in terms of mean square error (MSE).…”

“…Both papers present closed-form solutions on the tracking performance in terms of mean square error (MSE). Particularly, Zakia, et al [6] showed that a constant LOS component gives advantage in tracking the channel. This means that the train suffers larger tracking loss as it travels to a lower elevation angle.…”

“…However, their method is based on a terrestrial system that assumes the channel as frequency-selective so that it employs the weighted average of the Doppler shift of the channel paths, as opposed to a single path Doppler shift compensation as performed here. Moreover, our proposed method is motivated by the analytical work of Zakia, et al [6], which opens up the possibility of extending the result analytically. The performance of the proposed method was analyzed in terms of MSE and bit error rate (BER).…”

“…Secondly, even though we assume a single user SIMO beamforming framework similar to the method shown in White, et al [4], Zakia, et al [5], and Rappaport [11], the generalization to the MIMO case is straightforward as long as the Doppler shift of all channel pairs is assumed to be equal. Finally, the purpose of the adaptive filter is not only beamforming, as assumed here, but also channel tracking, like the one used in Zakia, et al [6].…”