Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.484872
Title: Long-range prediction of wireless channels and transmit preprocessing techniques
Author: Liu, Wei
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2007
Availability of Full Text:
Full text unavailable from EThOS.
Please contact the current institution’s library for further details.
Abstract:
The advanced channel quality-aware adaptive modulation and coding techniques employed in both existing and future wireless communication systems are capable of substantially improving the achievable system performance. Furthermore, other novel techniques employed by the base-station (BS), such as transmit preprocessing, pre-equalization and so on can be used for simplifying the design of the receiver. All these techniques require accurate channel state information (CSI) at the transmitter side, which necessutate the channel prediction. Both minimum mean square error (MMSE) channel predictors and Kalman-filtering assisted channel predictors are investigated in the context of narrowband channels. Then, for wideband channel, two-dimensional (2D) channel estimation and prediction was considered, which was capable of predicting both the frequency-domain (FD) and time-domain (TD) fluctuation of wideband channels. An eigenmode transmission based single-user multiple input multiplt output (MIMO) system was also investigated, which requires the computation of the singular vectors. These can be determined from the singular value decomposition (SVD) of the channel's impulse response (CIR) matrix, which has to be carried out at the regular instants and hence imposes a high computational complexity. However, instead of the periodic estimation of the CIR matrix and its regualr SVD, it is possible to directly track the output of the SVD, namely the sigular vectors without performing the abovementioned channel estimation and SVD. As far as MIMO aided multi-user systems are concerned, both zero-forcing and MMSE BS preprocessing techniques were investigated, which aim for simplifying the design of the mobile station's (MS) receiver. Again, channel prediction was invoked for acquiring the CSI required for transmit preprocessing. Furthermore, a SVD based transmit preprocessing algorithm was proposed for both uplink (UL) and downlink (DL) transmissions in the context of a MIMO system supporting multiple users and different power allocation schemes are designed for both UL and DL transmissions. Finally, the thesis was concluded with the investigation of recurrent neural network (RNN) based nonlinear channel prediction.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.484872  DOI: Not available
Share: