Search

We derive a new iterative multiuser detection algorithm based on a generalized sidelobe canceller which uses the constant modulus property of digitally modulated signals together with code knowledge given to a receiver in a CDMA system. The proposed algorithm can be seen as an extension of the least squares constant modulus algorithm in a sense that it simultaneously minimizes the constant modulus (CM) and the least square (LS) cost functions iteratively. The proposed detector performs as well as the ideal MMSE detector in the high SNR region, in addition to resolving the interference capture problem existing in most constant modulus algorithms (CMA).

We consider practical two-way communication schemes with blind channel estimation in TDD (time-division duplex) MISO (multi-input single-output) and MIMO (multi-input multi-output) systems. In the MISO system, we show that blind channel estimation only needs to be done at a transceiver equipped with multiple antennas and this estimated channel can be efficiently utilized for both uplink and downlink communication. In the MIMO system, we propose a two-stage communication scheme which begins with uplink transmission using differential space-time block codes and follows with two-way beamforming with blind channel estimation. We demonstrate, through simulations, that the novel schemes possess very good performance and compare them with training-based schemes.

The desire to reduce complexity at the mobile unit in cellular systems has led to the investigation of signal processing techniques that move computational complexity from the mobile unit to the base station. In this paper we investigate transmitter precoding for downlink time diversion duplex (TDD) code-division multiple-access (CDMA) communications. In particular, we develop a linear MMSE-based precoding strategy using blind channel estimation for fading multipath channels that allows for simple matched filtering at the mobile unit and is easy to make adaptive. We also present a performance analysis using tools developed for the analysis of conventional (receiver-based) linear blind multiuser detection in unknown channels. We compare the analytical and simulation results to traditional receiver-based blind multiuser detection. It is seen that transmitter precoding offers a reasonable alternative for TDD-mode CDMA when minimizing computational complexity at the mobile unit is a priority.

Smart antenna and associated technologies are expected to play a significant role in enabling broadband wireless communication systems. Smart antennas exploit space diversity to help provide high data rates, increased channel capacity, and improved quality of service at an affordable cost. In this paper we present a new procedure for implementing smart antenna algorithms. It is a hybrid approach that integrates the features of the switched beam method and the adaptive beam forming approach. Specifically it is shown that by using high gain antenna elements and combining the switched beam process with the adaptive beam forming procedure on a limited number of elements (as low as two in an eight-element array), a performance close to that of a more complex eight-element adaptive array may be achieved. The proposed hybrid method, therefore, is fast, is computationally efficient, and provides a cost effective approach for exploiting space diversity. Even with the inclusion of interference signals, the proposed hybrid approach out-performed the switched beam method, and provided performance similar to that of an adaptive array with less number of elements (three in an eight-element array). Implementation of an adaptive array also includes estimations; hence, reducing the number of elements in an array may lead to improved accuracy, in addition to fast convergence and reduced complexity.

We describe an approach to the performance analysis of the recently developed group-blind multiuser detector with known channels and blindly estimated channels. Specifically, we get the asymptotic distribution of the estimated group-blind detectors with known channels, and numerical ground to obtain an analytical Hermitian matrix with the blindly estimated channels. These make it possible to express the signal to interference plus noise ratio (SINR) analytically. Analytical results obtained in this paper match very well with the simulation results.

In blind or group-blind linear multiuser detection, the detector is estimated directly from the received signals, based on some partial knowledge about the system. The basic scenario is that the receiver has only limited knowledge about the CDMA channel, e.g., the signature waveform of a given user (or the signature waveforms of some but not all users group-blind detection), based on which a linear multiuser detector is estimated from the received signals and subsequently employed to demodulate the data symbols of the given user. In this paper, we address the following fundamental problem: Given a particular linear multiuser detector, e.g., the linear MMSE detector, and given certain prior knowledge about the channel, what is the best achievable performance of the corresponding blind or group-blind implementation of this linear detector, in terms of the output signal-to-interference-plus-noise ratio (SINR)?.

We consider wireless relaying: one or more nodes in a wireless (ad-hoc) network assist other nodes in their transmission by partially retransmitting messages. A characteristic of wireless relays - as compared to the work by T.M. Cover and A.A. El Gamal on the relay channel (see IEEE Trans. on Inf. Theory, vol.25, no.5, p.572-84, 1979) - is that they cannot transmit and receive simultaneously at the same frequency. We derive new upper and lower bounds for the capacity of this wireless relay channel. We then apply these result to a 4 terminal network, and show that the gain (considering outage capacity) of using wireless relaying is of the order of 8-9 dB.

This paper considers DS-CDMA systems with both inter- and intra-cell interference. The users inside the cell are assumed to have known codes, while the codes of the users outside the cell are unknown. To reduce the interference from both known and unknown users, a hybrid group-blind multiuser detector, derived in a previous paper, can be used. In the current paper we consider real time implementation using subspace tracking, and adapt the NA-CSVD (noise average cross-terms singular value decomposition) to the hybrid group-blind multiuser detector. The NA-CSVD algorithm is a new efficient QR Jacobi-type subspace tracking algorithm which has complexity O(Nd), where N is the processing gain and d is the number of out-of-cell users. It is shown that the NA-CSVD implementation of the hybrid multi-user detector has good performance, and that the performance considerably exceeds that of the purely blind detector, because of the sensitivity of the NA-CSVD to subspace dimension and the lower subspace dimension of the group-blind detector.

Previously developed blind techniques for multiuser detection in code division multiple access (CDMA) systems lead to several near-far resistant adaptive receivers for demodulating a given user's data with the prior knowledge of only the spreading sequence of that user. In the CDMA uplink, however, typically the base station receiver has the knowledge of the spreading sequences of all the users within the cell, but not that of the users from other cells. In this paper, group-blind techniques are developed for multiuser detection in such scenarios. These new techniques make use of the spreading sequences and the estimated multipath channels of all known users to suppress the intracell interference, while blindly suppressing the intercell interference. Several forms of group-blind linear detectors are developed based on different criteria. Moreover, group-blind multiuser detection in the presence of correlated noise is also considered. In this case, two receiving antennas are needed for channel estimation and signal separation. Simulation results demonstrate that the proposed group-blind linear multiuser detection techniques offer substantial performance gains over the blind linear multiuser detection methods in a CDMA uplink environment.

The recently developed blind adaptive techniques for multiuser detection in code division multiple access (CDMA) systems offer an attractive compromise of performance and complexity. However, the desire to further reduce complexity at the mobile unit has led to the investigation of techniques that move signal processing from the mobile unit to the base station. In this paper, we investigate transmitter precoding for downlink time division duplex (TDD) code division multiple access (CDMA) communications. In particular, we develop a linear minimum mean square error precoding strategy using blind channel estimation for fading multipath channels that allows for simple matched filtering at the mobile unit and is easy to make adaptive. We also present a performance analysis using tools developed for the analysis of conventional (receiver-based) linear blind multiuser detection in unknown channels. We compare the analytical and simulation results to traditional receiver-based blind multiuser detection. It is seen that transmitter precoding offers a reasonable alternative for TDD-mode CDMA when minimizing computational complexity at the mobile unit is a priority.

In a cooperative diversity network users cooperate on transmitting each others' messages; to some extend nodes therefore collectively act as an an- tenna array. In this paper we derive upper and lower bounds for the information theoretic capacity of a 4 node ad-hoc network with two transmitters using co- operative diversity.

We consider the problem of multi-user detection for CDMA systems where the codes of some users are known while others are unknown, semi-blind multi-user detection. An example is at the base station of a cellular communication system with interference from both in-cell users, with known codes and out-of-cell users, with unknown codes. We develop a number of semi-blind multi-user detectors using a subspace approach. One class of detectors developed has a computationally efficient structure of a subspace estimation common to all known users, followed by a simple processing specific to each user. The performance of the detectors is compared to that of the purely blind detector and the non-blind detector and the semi-blind detectors are seen to have the best performance.

This paper analyses combinations of the linear MMSE detector and a non-linear parallel interference canceller (PIC) for multi-user CDMA detection. Approximate formulas and upper and lower bounds for the error probability are derived and verified through simulation. It is shown that the combined MMSE/PIC detector almost always is better than the MMSE, and a partial MMSE/PIC is derived that has superior performance in all cases. It is also shown that although the MMSE detector is near-far resistant, it is sensitive to the code cross-correlation matrix, while the combined MMSE/PIC detector is both near-far resistant and insensitive to the cross-correlation matrix.

This paper analyses combinations of the linear MMSE detector and a non-linear parallel interference canceller (PIC) for multi-user CDMA detection. Both the case where all users' codes are known at the receiver, and the case where only some codes are known are considered. An upper bound and an approximate formula for the error probability are derived and verified through simulation It is shown that the combined MMSE/PIC detector can have a considerable performance gain over the MMSE. It is also shown that while the MMSE detector is sensitive to large code cross-correlation values, the combined MMSE/PIC detector is robust to code cross-correlations. Finally, use of the MMSE/PIC detector for cellular systems with both inter- and intra-cell interference is considered.

We consider the problem of multi-user detection for CDMA systems where the codes of some users are known while others are unknown, semi-blind multi-user detection. An example is at the base station of a cellular communication systems with interference from both in-cell users, with known codes, and out-of-cell users, with unknown codes. We develop a number of decorrelating, semi-blind multi-user detectors, using a subspace approach. One of the detectors has a computationally efficient structure of a subspace estimation common to all known users, followed by a simple processing specific to each user. The performance of the detectors is compared to that of the purely blind decorrelating detector and the non-blind decorrelating detector, and the semi-blind detectors are seen to have the best performance.

The effect of sampling and quantization on frequency estimation for a single sinusoid is investigated. The Cramer-Rao bound for 1 bit quantization is derived, and compared with the limit of infinite quantization. It is found that 1 bit quantization gives a slightly worse performance, however, with a dramatic increase of variance at certain frequencies. This can be avoided by using 4 times oversampling. The effect of sampling when using non-ideal antialiasing lowpass filters is therefore investigated. Cramer-Rao lower bounds are derived, and the optimal filters and sampling frequencies are found. Finally, fast estimators for 1 bit sampling, in particular correlation based estimators, are derived. The paper concludes with simulation results for 4 times oversampled 1 bit quantization.

In blind multiuser detection for CDMA systems, the receiver knows only the code of the user of interest, while in group-blind multiuser detection the receiver knows a subset of codes, e.g., the in-cell users in a base station. This paper derives bounds for the performance of linear blind and group-blind multiuser detectors. The bounds are derived under a number of different system assumptions. The bounds show the theoretical gain by using group-blind detectors over blind detectors, and they also show that previously developed blind and group-blind detectors are relatively close to optimum among estimators using only second order moments. However, the bounds also show that a considerable gap exists to optimum detectors that are not restricted to second order moments.

In blind and group-blind linear multiuser detection, the detectors are estimated from the received signals. The performance of a number of estimated linear detectors, including the direct-matrix-inversion (DMI) blind linear MMSE detector, the subspace blind linear MMSE detector, the form-I and form-II group-blind linear hybrid detectors, is analyzed. Asymptotic limit theorems for each of the estimates of these detectors (when the signal sample size is large) are established, based on which the average output signal-to-inference-plus-noise ratios (SINRs) of these detectors are given.

Due to the asymmetric allocation of uplink and downlink time slots in Universal Mobile Telecommunications System Terrestrial Radio Access Time Division Duplex (UTRA-TDD), the performance of mobile stations can be degraded seriously by inter-cell interference. We propose the group-blind hybrid multiuser detector for UTRA-TDD, which can reduce intra-cell and inter-cell interference efficiently. The detector has better performance compared to conventional multiuser detectors in the case of presence of unknown inter-cell interference. The group-blind hybrid multiuser detector can mitigate the problem of UTRA-TDD efficiently.

This paper considers the real-time implementation of group-blind multiuser detection in the uplink of CDMA systems. The class of group-blind multiuser detectors is used in the situation where a group of users have unknown codes, while another has known codes, for example in the uplink of a CDMA system. Previous papers have only considered block based implementation. In this paper, for the real-time implementation and tracking the non-stationary channel characteristics, we propose using the FASIR (FAst Subspace Iteration with Ritz acceleration) algorithm for subspace tracking. Moreover, we use a Kalman-type of algorithm for the blind channel estimation.

Macroscopic antenna diversity (macrodiversity) is a form of diversity where the branch signals originate from multiple base stations. While some macrodiversity techniques have been employed in modern cellular systems, techniques that employ combining have not been explored as a means of reducing interference in these systems. In CDMA, the use of more sophisticated techniques could significantly reduce multiple access interference and extend the capacity of current and next-generation systems. A novel combining technique based on macrodiversity is presented. The proposed system has been shown to provide significant capacity improvements in simulation, and shows significant promise as a cost-effective upgrade option to cellular service providers.

We present some key advanced signal processing methodologies that have been developed for interference suppression in wireless networks. We focus primarily on the problem of jointly suppressing multiple-access interference (MAI) and intersymbol interference (ISI), which are the limiting sources of interference for the high data-rate wireless systems being proposed for many emerging application areas, such as wireless multimedia. We first present a signal subspace approach to blind joint suppression of MAI and ISI. We then discuss a powerful iterative technique for joint interference suppression and decoding, the so-called turbo multiuser detection, that is especially useful for wireless multimedia packet communications. We also discuss space-time processing methods which employ multiple antennas for interference rejection and signal enhancement. Finally, we touch on the problems of suppressing narrowband interference and impulsive ambient noise, two other sources of radio frequency interference present in wireless multimedia networks.

This paper considers the problem of multiuser detection for CDMA systems where the codes of a group of users are known while others are unknown, called group-blind multiuser detectors. An example is the uplink in cellular communication system with interference from both in-cell users, with known codes, and out-of-cell users, with unknown codes. In previous papers linear detectors for this situation were developed. In the current paper, we combine these methods with low-complexity non-linear post processing, using the method of slowest descent, a general method for non-linear optimization. The performance of this new group-blind detector is compared to that of the non-blind MMSE detector as well as the linear group blind detector, and the non-linear detector is seen to have a considerably better performance, in particular in heavily loaded systems.

In this paper, blind and group-blind detectors are proposed based on the constant modulus (CM) property of the transmitted signals. By adopting blind and group-blind CM detectors, signi£cant gains are attained compared to the original blind and group-blind detectors, with only a slight increase in complexity. The SINR of the blind and group-blind CM detectors approach the theoretical upper bound for SINR (Cramer-Rao bound). Furthermore, the proposed detectors perform almost as well as the exact MMSE detector at high SNR.

Blind channel estimation is the process of obtaining the impulse responses between a source and multiple (arbitrarily placed) receivers without prior knowledge about the source characteristics or the environment. This approach could simplify localization of non-impulsive submerged sound sources (e.g., pinnipeds or cetaceans); the process of picking arrivals (direct and reflected) could be carried out on the estimated impulse responses rather than on the recorded waveforms, thus facilitating the use of time of arrival-based localization approaches. Blind channel estimation could also be useful in estimating the original source signal of a vocalizing animal through deconvolution of the estimated channel impulse responses and the recorded waveforms. In this paper, simulation and controlled pool studies will be used to explore requirements on source and environment characteristics and to quantify blind channel estimation performance for underwater passive acoustic applications.

The effect of sampling and quantization on frequency estimation for a single sinusoid is investigated. Asymptotic Cramer-Rao bounds (CRB) for 1-bit quantization and for non-ideal filters are derived, which are simpler to calculate than the exact CRB while still relatively accurate. It is further investigated how many bits should be used in quantization to avoid the problems of 1-bit quantization, and it turns out that 3-4 bits are enough. Finally, oversampled 1-bit quantization is investigated. It is determined how much the signal should be oversampled, and in addition /spl Sigma//spl Delta/ modulators are investigated.

We consider the problem of multi-user detection for CDMA systems where the codes of some users are known while others are unknown, called semi-blind detectors. An example is at the base station of a cellular communication system with interference from both in-cell users, with known codes, and out-of-cell users, with unknown codes. We develop a hybrid semi-blind detector, which is partly decorrelating, partly MMSE. Subspace tracking methods are developed for on-line implementation of the detector. The performance of the detectors is compared to that of the purely blind MMSE detector and the non-blind MMSE detector, and the semi-blind detector is seen to have a considerable better performance.

We developed a method to separate click trains from multiple odontocetes that relies on click timing only. The method assumes that inter-click intervals (ICIs) are slowly varying and that their distribution is known, though possibly with unknown parameters. ICIs are modeled as a renewal process and click trains are separated by maximizing the likelihood of click assignments based on ICI distributions. We present results from application of this timing-based separation method to simulated and real datasets. [Work supported by the National Science Foundation and the Office of Naval Research.]

In this paper we evaluate the effect of quantization of Doppler signals. We do this by calculating Cramer-Rao lower bounds for one bit quantization and by doing computer simulations. Both approaches show that there are serious problems with one bit quantization around certain critical frequencies. Contrasting, it is shown that 4 bit quantization performs nearly as good as no quantization. The conclusion is that 4 bits are adequate while fewer number of bits give unreliable results.

Nonlinear precoding schemes for downlink time-division duplex-CDMA systems over multipath fading channels, are considered. First, the capacity results of a downlink CDMA system with either multiuser detection or precoding, were obtained and compared. It is seen that the two schemes exhibit similar capacity regions for both sum rate and maximum equal rate, which motivates the development of efficient nonlinear transmitter precoding techniques to reduce the receiver complexity at the mobile units without degrading the system performance. We then develop both bit-wise and chip-wise Tomlinson-Harashima (TH) multiuser precoding methods for downlink CDMA with multipath, to remove multi-user interference, inter-chip interference and inter-symbol interference. Efficient algorithms for multiuser power loading and ordering are also developed. Implementation of the proposed TH-precoding schemes in time-varying channels based on channel prediction is addressed as well. Simulations results are provided to demonstrate the effectiveness of the proposed techniques in suppressing interference in downlink CDMA.

Computer-simulation techniques are used to study the domain-growth kinetics of (2×1) ordering in a two-dimensional Ising model with nonconserved order parameter and with variable ratio α of next-nearest- and nearest-neighbor interactions. At zero temperature, persistent growth characterized by the classical growth exponent n≃1 / 2 is found for α≥1, whereas the domain boundaries become pinned and the growth stops for α

This paper provides an asymptotic performance evaluation of network-based or multilateral positioning systems, where the times and/or angles of arrival of the signals traveling from a mobile station (MS) to a set of fixed stations (FSs) are used to estimate the MS's position. The effects of the various system-design parameters as well as the effects of biased bearing and time measurements are analyzed, and explicit formulas are derived. Biased measurements can result, for example, from hardware calibration errors or multipath effects, including the lack of line of sight (LoS). Among other results, it is shown that only a few system and cellular parameters can be used to minimize their nuisance effects. The mathematical analysis is based on the Cramer-Rao lower bound (CRLB) and the maximum likelihood theory. It is completed with numerical simulations and a discussion of the positioning performance for two important second-generation cellular systems, GSM and IS-95.

Commercial applications for the location of subscribers of wireless services continue to expand. Consequently, finding the Cramer-Rao lower bound (CRLB), which serves as an optimality criterion for the location estimation problem, is of interest. In this paper, we derive the deterministic CRLBs for the estimation of the specular multipath parameters and the positions of the mobiles in an asynchronous direct sequence code division multiple access (DS-CDMA) system operating over specular multipath fading channels. We assume a multilateral radio location system where the location estimates are obtained from some or all of the estimated signal parameters at different clusters of antennas of arbitrary geometry. Extension for unilateral and composite radio location techniques is also discussed. As an application example, we use numerical simulations to investigate the effects of specular multipath and multiple access interference (MAI) on the positioning accuracy for different radio location techniques.

This paper provides an asymptotic performance evaluation of network based or multilateral radio location estimators that use the time of arrival (TOA) and/or the angle of arrival (AOA) measurements of the signals travelling from a mobile station (MS) to a set of fixed stations (FSs) to estimate the MS's position. The analysis is based on the Cramer-Rao lower bound (CRLB) and the maximum likelihood theory, and takes into account the effects of the system design parameters as well as the effects of biased measurements, which can result for example from hardware calibration errors or multipath effects, including the lack of line of sight. Explicit formulas that provide a clear relationship between the various system design parameters and the positioning accuracy are derived and discussed. Finally, the theory is verified with numerical simulations.

Commercial applications for the location of subscribers ofwireless services continue to expand. Consequently, 'nding the Cramer- Rao bound (CRB), which serves as an optimality criterion for the location estimation problem, is of interest. In this article, we derive the CRB for the estimation of channel parameters and mobile position in an asynchronous direct sequence code division multiple access (DS-CDMA) system operating over fading channels. It is assumed that the location estimates are obtained from the bearings and/or propagation delays estimated at one or more cluster(s) of antenna array of arbitrary geometry. Among other applications, the CRB on the positioning accuracy may serve as a design tool to 'nd an optimum antenna placement, or to evaluate the practicability of a legal demand for emergency location.

A mathematical treatment of particle size and velocity estimation from two channel laser anemometry measurements is considered. Cramer-Rao bounds for the general case are derived, and the corresponding maximum likelihood estimator is analyzed through computer simulations. Low complexity correlation based estimators are derived and their performance is characterized. The results predicted by theory are illustrated by some numerical examples.