Channel tracking for RAKE receivers in closely spaced multipathenvironments

This paper deals with the problem of channel tracking for RAKE receivers in propagation environments characterized by closely spaced multipath components. After outlining why conventional single-path channel tracking algorithms fail in such scenarios, several new estimation algorithms are developed that are tailored to channels with closely spaced multipaths. This is achieved by removing or minimizing self-interference caused by multipath components. Other interfering users are treated as noise. Both timing tracking and phasor tracking and their interaction are covered in this paper. The derived algorithms are benchmarked against perfect channel knowledge on one hand and conventional tracking algorithms on the other hand, both in a UMTS test scenario. In moderate scenarios, the use of these new algorithms leads to performance improvements of up to 2 dB, in terms of signal-to-noise ratio (SNR) at moderate bit error rates, and even manages to track the channel in conditions where conventional tracking algorithms fail completely     

Achievable rate of MIMO channels with data-aided channel estimationand perfect interleaving

The achievable rate of a coherent coded modulation digital communication system with data-aided channel estimation and a discrete equiprobable symbol alphabet is derived under the assumption that the system operates on a flat fading multiple-input/multiple-output channel and uses a perfect interleaver to combat the bursty nature of the channel. It is shown that linear minimum mean square error channel estimation directly follows from the derivation and links average mutual information to the channel dynamics. Based on the assumption that known training symbols are transmitted, the achievable rate of the system is optimized with respect to the amount of training information needed     

An information theoretic foundation of synchronized detection

The constrained capacity of a coherent coded modulation (CM) digital communication system with data-aided channel estimation and a discrete, equiprobable symbol alphabet is derived under the assumption that the system operates on a flat fading channel and uses an interleaver to combat the bursty nature of the channel. It is shown that linear minimum mean square error channel estimation directly follows from the derivation and links average mutual information to the channel dynamics. Based on the assumption that known training symbols are transmitted, the achievable rate of the system is optimized with respect to the amount of training information needed. Furthermore, the results are compared to the additive white Gaussian noise channel, and the case when ideal channel state information is available at the receiver     

Nonlinear adaptive prediction of speech with a pipelined recurrentneural network

New learning algorithms for an adaptive nonlinear forward predictor that is based on a pipelined recurrent neural network (PRNN) are presented. A computationally efficient gradient descent (GD) learning algorithm, together with a novel extended recursive least squares (ERLS) learning algorithm, are proposed. Simulation studies based on three speech signals that have been made public and are available on the World Wide Web (WWW) are used to test the nonlinear predictor. The gradient descent algorithm is shown to yield poor performance in terms of prediction error gain, whereas consistently improved results are achieved with the ERLS algorithm. The merit of the nonlinear predictor structure is confirmed by yielding approximately 2 dB higher prediction gain than a linear structure predictor that employs the conventional recursive least squares (RLS) algorithm