Damien Roque

Performances of Weighted Cyclic Prefix OFDM with Low-Complexity Equalization

In this paper, we justify low-complexity equalization techniques for weighted cyclic prefix (WCP)-OFDM. This modulation technique refers to filter bank based multicarrier (FBMC) transmission system provided with short filters. It allows the use of non-rectangular waveforms in order to mitigate interference caused by time-frequency selective channels while preserving an efficient implementation.

A performance comparison of FBMC modulation schemes with short perfect reconstruction filters

In this paper, we study the performance of filter bank based multicarrier (FBMC) transmission systems over doubly-dispersive channels. FBMC generalizes traditional orthogonal frequency-division multiplexing (OFDM) schemes, allowing a non-rectangular sub-channel pulse shape in the time domain. This approach leads to a better spectral containment that improves interference mitigation in several time-variant environments. A general analysis describes the transmultiplexer structure and the time-variant multipath channel model. We specify rectangular filters and short perfect reconstruction (PR) filters that give rise to two families of FBMC transceivers. A performance comparison between these schemes is performed in terms of bit-error-rate using a single-tap per sub-channel equalizer. This study emphasizes the low complexity of the FBMC transceivers in the case of short filters and discusses its usefulness through simulation results in two mobility scenarios.

On the study of faster-than-Nyquist multicarrier signaling based on frame theory

Multicarrier transmissions are classically based on undercomplete or exact Weyl-Heisenberg Riesz (biorthogonal or orthogonal) bases implemented thanks to oversampled filter-banks. This can be seen as a transmission below the Nyquist rate. However, when overcomplete Weyl-Heisenberg frames are used, we obtain a “faster-than-Nyquist” (FTN) system and it is theoretically impossible to recover exactly transmitted symbols using a linear receiver. Various studies have shown the interest of this high density signaling scheme as well as practical implementations based on trellis and/or iterative decoding. Nevertheless, there is still a lack of theoretical justifications with regard to pulse design in the FTN case. In this paper, we consider a linear transceiver operating over an additive white Gaussian noise channel. Using the frame theory and simulation results, we show that the mean squared error (MSE) is minimized when tight frames are used.

Weighted cyclic prefix OFDM: PAPR analysis and performances comparison with DFT-precoding

In this paper, we present a weighted cyclic prefix orthogonal frequency-division multiplexing (WCP-OFDM) transceiver as a generalization of traditional cyclic prefix (CP)-OFDM. In time-variant channels, this multicarrier transmission scheme may mitigate inter-channel interference (ICI) thanks to the use of non-rectangular pulse shapes. A precoding step may be required in order to reduce the peak-to-average power ratio (PAPR) at the transmitter output. For instance, a discrete Fourier transform (DFT) precoder leads to a single carrier transmission scheme with frequency domain equalization. We analyze the consequences of such a precoding, in terms of performances, in the context of a time-frequency selective channel.

Non-rectangular perfect reconstruction pulse shaping based ICI reduction in CO-OFDM

In this paper, we propose to increase residual carrier frequency offset tolerance based on short perfect reconstruction pulse shaping for coherent optical-orthogonal frequency division multiplexing. The proposed method suppresses the residual carrier frequency offset induced penalty at the receiver, without requiring any additional overhead and exhaustive signal processing. The Q-factor improvement contributed by the proposed method is 1.6 dB and 1.8 dB for time-frequency localization maximization and out-of-band energy minimization pulse shapes, respectively. Finally, the transmission span gain under the influence of residual carrier frequency offset is ~62% with out-of-band energy minimization pulse shape.

Improving Spectral Efficiency While Reducing PAPR Using Faster-Than-Nyquist Multicarrier Signaling

Multicarrier modulations are widely used in mobile radio applications due to their adaptability to the time-frequency characteristics of the channel, thus enabling low-complexity equalization. However, their intrinsically high peak-to-average power ratio (PAPR) is a major drawback with regard to implementation issues (e.g., power amplification efficiency, regulatory constraints...).

A low-complexity multicarrier scheme with LDPC coding for mobile-to-mobile environment

In this paper, we consider a mobile ad-hoc radio network in an urban area. The propagation environment between two endpoints can be modeled by a double-Rayleigh fading multipath channel. Such a mobile scenario justifies the use of filter bank based multicarrier (FBMC) transmission systems. This technique generalizes traditional cyclic prefix orthogonal frequency-division multiplexing (CP-OFDM), allowing the design of non-rectangular pulse shape filters. We show that this approach leads to a better interference mitigation in time-variant channels. We restrict our study to short filters and single-tap per sub-channel equalization in order to preserve a low-complexity transmultiplexer. In this study, we compare FBMC with short filters to CP-OFDM in terms of coded bit-error-rate performances, using a realistic mobile-to-mobile channel model.

Closed-Form Expressions for Channel Shortening Receivers Using

Channel shortening has been studied in the context of intersymbol interference and multi-in multi-out channels as a means to compute a posteriori probabilities with a BCJR algorithm at a reduced computational complexity. This is done by considering an approximate channel response of reduced length. In a turbo receiver, soft a priori information can be linearly combined with the received sequence to form a new input to the BCJR trellis-based processing. In this letter, we provide closed-form expressions for the channel shortening filters using a generalized mutual information objective function. The proposed receiver allows a complexity reduction with respect to numerical optimization approaches which may also present stability, precision, and convergence issues.

Blind Symbol Rate Estimation of Faster-than-Nyquist Signals Based on Higher-Order Statistics

Both faster-than-Nyquist (FTN) and cognitive radio go towards an efficient use of spectrum in radio communications systems at the cost of an added computational complexity at the receiver side. To gain the maximum potential from these techniques, non-data-aided receivers are of interest. In this paper, we use fourth-order statistics to perform blind symbol rate estimation of FTN signals. The estimator shows good performance results for moderate system’s densities beyond the Nyquist rate and for a reasonable number of received samples.

Analysis of a FTN Multicarrier System: Interference Mitigation Based on Tight Gabor Frames

Cognitive radio applications require flexible waveforms to overcome several challenges such as opportunistic spectrum allocation and white spaces utilization. In this context, multicarrier modulations generalizing traditional cyclic-prefix orthogonal frequency-division multiplexing are particularly justified to fit time-frequency characteristics of the channel while improving spectral efficiency.In our theoretical framework, a multicarrier signal is described as a Gabor family the coefficients of which are the symbols to be transmitted and the generators are the time-frequency shifted pulse shapes to be used. In this article, we consider the case where non-rectangular pulse shapes are used with a signaling density increased such that inter-pulse interference is unavoidable. Such an interference is minimized when the Gabor family used is a tight frame. We show that, in this case, interference can be approximated as an additive Gaussian noise. This allows us to compute theoretical and simulated bit-error-probability for a non-coded system using a quadrature phase-shift keying constellation. Such a characterization is then used in order to predict the convergence of a coded system using low-density parity check codes. We also study the robustness of such a system to errors on the received bits in an interference cancellation context.