Showing papers by "Timothy A. Thomas published in 2002"
17 Mar 2002
TL;DR: Results indicate that CP-CDMA provides significantly improved performance, yet has lower complexity than the time-domain equalized system.
Abstract: This paper describes a novel type of single-carrier DS-CDMA system that combines the use of OFDM-style cyclic prefixes with frequency-domain equalization, for use on the forward link of broadband CDMA cellular systems. The proposed method is referred to as CP-CDMA. The combination of cyclic prefixes and MMSE frequency-domain equalization solves many of the performance and complexity problems that arise in CDMA receivers when techniques such as multi-level modulation and full code usage are employed. Simulation results are presented for a 20 MHz system with multi-level modulation and these results are compared to the performance of a conventional CDMA system with either a RAKE receiver or a time-domain equalizer. Results indicate that CP-CDMA provides significantly improved performance, yet has lower complexity than the time-domain equalized system.
138 citations
Patent•
18 Jul 2002TL;DR: In this article, a receiver for implementing a training prefix modulation method in response to a reception of a signal propagating through a channel is disclosed, and an estimate of the data inter-block interferences is generated and added to the data blocks.
Abstract: A receiver for implementing a training prefix modulation method in response to a reception of a signal propagating through a channel is disclosed. The signal as received includes training blocks with each training block having a data inter-block-interference therein, and data blocks with each data block having a training inter-block-interference therein. The signal is selectively reconstructed to provide a circular appearance of the channel over the data blocks. Specifically, an estimate of the training inter-block-interferences is generated and subtracted from the data blocks. And, an estimate of the data inter-block interferences is generated and added to the data blocks.
74 citations
07 Aug 2002
TL;DR: Simulation results show that the proposed transmit diversity technique and receive frequency domain equalization provide vastly superior performance than previous Walsh-code-based transmit diversity methods using RAKE reception.
Abstract: This paper explores the addition of antenna diversity schemes to the recently-proposed cyclic-prefix CDMA (CP-CDMA) system concept CP-CDMA is a novel type of single carrier DS-CDMA that employs OFDM-style cyclic prefixes for the purposes of facilitating efficient frequency-domain equalization in broadband CDMA cellular systems. CP-CDMA is intended for high-bandwidth, high-delay-spread cellular channels (especially in the forward link), such as those envisioned for systems that evolve from current 3G CDMA standards. This paper proposes transmit and receive antenna diversity techniques for CP-CDMA systems that operate with high user loading and multi-level modulation in broadband high delay-spread channels. Simulation results show that the proposed transmit diversity technique and receive frequency domain equalization provide vastly superior performance than previous Walsh-code-based transmit diversity methods using RAKE reception.
60 citations
07 Aug 2002
TL;DR: The results include path loss measurements and various metrics of the MIMO channel, such as channel impulse responses, rms delay spread, and other spatial characteristics.
Abstract: This paper summarizes the propagation characteristics of a 20 MHz channel at a carrier frequency of 3.676 GHz transmitted from a two-antenna base to a two-antenna mobile receiver. The experiments were conducted in a moderate density suburban setting outside Chicago, Illinois. The transmit and receive systems are described, as are the data collection procedures and data analysis. The results include path loss measurements and various metrics of the MIMO channel, such as channel impulse responses, rms delay spread, and other spatial characteristics.
41 citations
13 May 2002
TL;DR: A training sequence is presented that enables other ToA estimators such as ESPRIT to determine multiple data stream's TOAs and shows that the new TOA estimator has the best overall performance.
Abstract: This paper contains a new frequency-domain multi-stream channel estimation technique for communication systems such as MIMO-OFDM. The estimator works by first determining the Time Of Arrivals (TOAs) of each data stream using a near Maximum Likelihood (ML) approach and then solving for the corresponding gains. In addition, this paper presents a training sequence that enables other TOA estimators such as ESPRIT to determine multiple data stream's TOAs. The performance of various channel estimators is compared through simulations and experimental data and the results show that the new TOA estimator has the best overall performance.
16 citations
10 Dec 2002
TL;DR: In this article, the authors proposed a power weighting scheme to improve the performance of a successive cancellation receiver in frequency-selective MIMO communications with equal power on each stream.
Abstract: Bell Labs layered space-time (BLAST) promises to increase greatly the capacity of future broadband mobile communications. However, in MIMO communication systems, the decoded BER performance of successive cancellation receivers is limited by the performance of the stream with the highest mean squared error. The paper presents a novel method to determine the different power weighting to be applied by the transmitter to each BLAST data stream in frequency-selective communications to improve the performance of a successive cancellation receiver. In addition, the new method predetermines the successive cancellation decoding order, thus simplifying the receiver. Simulation results for a spread OFDM system show up to a 5.0 dB improvement in frequency-selective channels over BLAST with equal power on each stream.
13 citations
10 Dec 2002
TL;DR: Examining both single-data stream and multiple data stream techniques for CP-CDMA systems that operate with high code usage and multi-level modulation in broadband high delay-spread channels for frequency-selective MIMO channels is examined.
Abstract: This paper explores the addition of frequency-domain transmit antenna arrays to the cyclic-prefix CDMA (CP-CDMA) system concept. CP-CDMA is a novel type of single carrier DS-CDMA that employs OFDM-style cyclic prefixes to facilitate efficient frequency-domain receiver processing in broadband CDMA cellular systems. CP-CDMA is intended for high-bandwidth, high-delay-spread cellular channels, such as those envisioned for systems that evolve from current 3G CDMA standards. This paper examines both single-data stream (i.e., maximal ratio transmission) and multiple data stream (i.e., spatial multiplexing also known as multiple-input, multiple-output or MIMO) techniques for CP-CDMA systems that operate with high code usage and multi-level modulation in broadband high delay-spread channels. In the proposed algorithms, both the transmit and receive array processing is performed completely in the frequency-domain to allow the tracking of frequency-selective MIMO channels. Simulation results are presented to show the efficacy of the proposed methods.
11 citations
04 Aug 2002
TL;DR: A frequency-domain channel estimation algorithm for single-user orthogonal frequency division multiplex (OFDM) wireless systems in the presence of interference that outperforms the least-squares method in accuracy, and requires a smaller number of pilots than the MANOVA method and thus has smaller overhead.
Abstract: We develop a frequency-domain channel estimation algorithm for single-user orthogonal frequency division multiplex (OFDM) wireless systems in the presence of interference. The received measurement is correlated in space with a covariance matrix dependent on frequency. Hence, the commonly used least-squares algorithm is suboptimal. On the other hand, accurate estimation of the spatial covariance matrix in such a model using the multivariate analysis of variance (MANOVA) method would impose significant computational overhead, since it would require a large number of pilot symbols. To overcome these problems, we propose to model the covariance matrix using an a-priori known set of frequency-dependent functions of joint (global) parameters, resulting in a structured covariance matrix. We estimate the interference covariance parameters using a residual method of moments (RMM) estimator and the mean (user channel) parameters by maximum likelihood (ML) estimation. Since our RMM estimates are invariant to the mean, this approach yields simple non-iterative estimates of the covariance parameters while having optimal statistical efficiency. Therefore, our algorithm outperforms the least-squares method in accuracy, and requires a smaller number of pilots than the MANOVA method and thus has smaller overhead. Numerical results illustrate the applicability of the proposed algorithm.
7 citations
17 Nov 2002
TL;DR: This paper considers practical link designs that include different transmit, receive, and coding schemes for an OFDM system with two transmit and two receive antennas and compares link designs in terms of simulated frame error rate (FER) performance in different propagation environments.
Abstract: This paper considers practical link designs that include different transmit, receive, and coding schemes for an OFDM system with two transmit and two receive antennas. In conjunction with studies on theoretical outage capacity, we compare link designs in terms of simulated frame error rate (FER) performance in different propagation environments. For channel coding, serial and parallel concatenated turbo-like codes are used and compared. For receive schemes, we compare maximum likelihood (ML) and linear MIMO receive algorithms and propose modifications to improve turbo decoding. For transmit schemes, we compare the Alamouti (1998) diversity scheme, a single-stream MIMO, and a multi-stream MIMO design. Interestingly, a 2/spl times/2 MIMO-OFDM design with ML receivers is found to under-perform the 2/spl times/2 Alamouti-OFDM design at low spectral efficiencies (e.g., 2, 4 b/subcarrier), even though MIMO has a slightly higher theoretical capacity. Also, the gap between the SNR required by the theoretical 1% outage capacity and that required by the single-stream MIMO-ML design at 10/sup -2/ FER is only 1-2 dB at spectral efficiencies of 2, 4 and 6 b/subcarrier, in an environment with rich spatial and frequency diversity.
5 citations
07 Aug 2002
TL;DR: A novel semi-blind multi-user channel estimator for OFDM based on the maximum likelihood estimator that is effective even when the frequency-domain pilot spacing is greater than the Nyquist spacing.
Abstract: Semi-blind channel estimators promise to dramatically reduce the number of pilots needed for accurate channel estimation over pilot-only based techniques while providing much better performance than blind algorithms. This paper presents a novel semi-blind multi-user channel estimator for OFDM based on the maximum likelihood estimator. The estimator works by iterating between estimating the channels between multiple transmit antennas and each receive antenna and the respective symbols for each transmitter. Simulation results show that the semi-blind channel estimator is effective even when the frequency-domain pilot spacing is greater than the Nyquist spacing.
4 citations