Zhengyuan Xu

A subspace approach to blind estimation of ultra-wideband channels

In impulse radio based ultra-wideband (UWB) multiuser communication systems, each users transmitted symbols are modulated by pulse position modulation (PPM), and corrupted by both intersymbol interference and multiple access interference at the receiver end. To construct low-complexity linear receivers for efficient symbol detection, channel information is needed. In this paper, following a particular input/output model that transforms a PPM system model into a sum of seemingly pulse amplitude modulation signals, a structure similar to a code-division multiple access system, we show that subspace technique can be applied to blindly estimate multipath channels. Performance of the proposed channel estimator is studied. To detect desired symbols, MMSE receivers are constructed. Simulation results demonstrate superiority of the proposed method to the existing maximum likelihood channel estimator.

Towards closing the gap between MOE and subspace methods

In this paper, we propose a power of R (POR) technique to significantly improve the performance of the minimum output energy (MOE) receiver. The new receiver is shown to asymptotically converge to the MMSE receiver. The convergence is established either under high SNR, or with large exponent raised in the power of the covariance matrix. Connection between the POR method and a subspace method is investigated, and their asymptotic equivalence is also established. However, the POR receiver does not require either a computationally expensive subspace decomposition of a data covariance matrix, or estimation of dimension of either the signal subspace or the noise subspace.

Perturbation study on MOE-based multiuser detection

The minimum output energy (MOE) receiver has been developed for multiuser detection. Its performance has been shown to be very close to the minimum mean square error (MMSE) receiver at high signal to noise ratio (SNR). However due to additive noise, the constraint vector required to construct the MOE receiver becomes a biased estimated channel vector. Thus the receiver is also perturbed. In addition to the noise, other factors affect the receivers performance as well. Effects from such factors as finite number of data samples and channel order overestimation are studied in this paper based on perturbation theory.

Blind channel estimation via subspace approximation

Subspace channel estimation technique has been well studied. It relies on subspace decomposition on the data covariance matrix to identify either signal subspace or noise subspace. However, the subspace decomposition process is computationally expensive. This paper applies a novel subspace approximation (SA) idea to bypass this process for channel estimation. It is based on a recently proposed power of R (POR) multiuser detection technique that raises power of estimated data covariance matrix to a positive integer to approximate rather than estimate the noise subspace component. Thus complexity is significantly reduced while satisfactory performance is still maintained. Channel estimation performance is studied in detail and compared with that of the well-known subspace method in the literature.

Subspace-based channel estimation for CDMA downlink with aperiodic spreading codes and multiple subchannels

We consider a downlink CDMA system with aperiodic spreading codes and multipath propagation. Assisted by either multiple antennas or oversampling, multiple subchannels are created. Then a subspace technique can be applied directly to estimate all channel parameters, based only on the channel output. The proposed method does not require either the spreading codes of the users or pre-processing of the received data. It is also applicable to an overloaded system. These appealing features make our method significantly different from existing subspace approaches.

Code constrained CMA-based multirate multiuser detection

In this paper, a constant modulus algorithm (CMA) based multiuser detector is proposed for multirate communications over unknown multipath channels. We impose multiple linear constraints on the detector to guarantee no cancellation of the desired signal, while suppressing the intersymbol interference (ISI) and multiuser interference (MUI). The constraint parameters can be either pre-selected or jointly optimized with the detector. Under certain conditions on the constraints, the proposed algorithm enjoys global convergence.

Blind detection of CDMA signals based on Kalman filter

This paper studies detection of CDMA signals based on the Kalman filter. In the absence of both training symbols and the knowledge about multipath channel parameters, the system can be exactly fitted into the Kalman model. By defining the combination of channel and transmitted signals as state variables, the state transition matrix and measurement matrix are derived without approximations. The covariance of process noise in the proposed algorithm is recursively estimated from the estimated states, in contrast to all existing approaches where channel parameters are assumed known or estimated by other methods. Therefore the proposed structure is much simpler and easy to implement.

A globally convergent CMA-based approach to blind multiuser detection

In this paper, we study blind multiuser detection in a CDMA system in the presence of multipath propagation. The receiver is forced to follow a minimum mean-square-error type parameterized by a channel like constraint vector. Then we minimize the Godards CMA cost function with respect to such a constraint vector in a much smaller dimensional space. Global convergence of the proposed approach is established. Performance of the channel estimator and receiver is also analyzed, together with comparisons with existing methods. Simulation results show that the proposed method performs well by both batch and adaptive implementations.

Improved constrained optimization method for CDMA systems

The constrained optimization method has been shown to be applicable for direct CDMA receiver design. However, due to the presence of additive noise and multipath distortion, the optimal constraint is a biased estimate for the channel, resulting in some performance loss of the constrained receiver compared with the MMSE receiver. This loss becomes more significant for the lower SNRs. It is shown in this paper that constrained optimization is closely related to the subspace method which can give a good estimate for the channel. Motivated by this result, a modified cost function is proposed to obtain the constraint with less complexity than the subspace method and better performance than the previously proposed constrained method.

Blind channel estimation for precoded variable bit-rate multiuser systems

The precoder has been shown to be able to provide source diversity and more freedom in design. In this paper we employ precoding techniques for block transmission based on a multirate filterbank structure. To meet multiuser communication requirements for various high data-rates, different filterbanks are used as precoders with corresponding coefficients and up/down sampling rates. However due to high speed communication in the presence of unknown multipath, different interferences exist in the received data such as multiuser interference, intersymbol interference and interblock interference. To estimate the multipath channel for a desired user we generalize the subspace method by employing all signatures associated with this channel in the current system. In this way a better performance for the channel estimate can be achieved. The delay for that user can also be jointly estimated together with channel estimation.