arXiv subjects

Amine Mezghani

Publications and source records attributed to Amine Mezghani.

78 records · Page 5Linked to original sources

arXiv2014

Information Theoretic Analysis of Concurrent Information Transfer and Power Gain

In this paper, we analyze the fundamental trade-off between information transfer and power gain by means of an information-theoretic framework in communications circuits. This analysis is of interest as many of today's applications require that maximum information and maximum signal power are extracted (or transferred) through the circuit at the same time for further processing so that a compromise concerning the signal spectral shape as well as the matching network has to be found. To this end, the optimization framework is applied to a two-port circuit, which is used as an abstraction for a broadband amplifier. Thereby, we characterize the involved Pareto bound by considering different optimization problems. The first one aims at optimizing the input power spectral density (PSD) as well as the source and load admittances, whereas the second approach assumes the PSD to be fixed and uniformly distributed within a fixed bandwidth and optimizes the source and load admittances only. Moreover, we will show that additional matching networks may help to improve the trade-off.

cs.IT↗

arXiv2014

A Lower Bound for the Fisher Information Measure

The problem how to approximately determine the absolute value of the Fisher information measure for a general parametric probabilistic system is considered. Having available the first and second moment of the system output in a parametric form, it is shown that the information measure can be bounded from below through a replacement of the original system by a Gaussian system with equivalent moments. The presented technique is applied to a system of practical importance and the potential quality of the bound is demonstrated.

cs.IT↗

arXiv2010

Analysis of 1-bit Output Noncoherent Fading Channels in the Low SNR Regime

We consider general multi-antenna fading channels with coarsely quantized outputs, where the channel is unknown to the transmitter and receiver. This analysis is of interest in the context of sensor network communication where low power and low cost are key requirements (e.g. standard IEEE 802.15.4 applications). This is also motivated by highly energy constrained communications devices where sampling the signal may be more energy consuming than processing or transmitting it. Therefore the analog-to-digital converters (ADCs) for such applications should be low-resolution, in order to reduce their cost and power consumption. In this paper, we consider the extreme case of only 1-bit ADC for each receive signal component. We derive asymptotics of the mutual information up to the second order in the signal-to-noise ratio (SNR) under average and peak power constraints and study the impact of quantization. We show that up to second order in SNR, the mutual information of a system with two-level (sign) output signals incorporates only a power penalty factor of almost 1.96 dB compared to the system with infinite resolution for all channels of practical interest. This generalizes a recent result for the coherent case.

cs.IT↗

arXiv2010

Achieving near-Capacity on Large Discrete Memoryless Channels

We propose a method to increase the capacity achieved by uniform prior in discrete memoryless channels (DMC) with high input cardinality. It consists in appropriately reducing the input set. Different design criteria of the input subset are discussed. We develop an efficient algorithm to solve this problem based on the maximization of the cut-off rate. The method is applied to a mono-bit transceiver MIMO system, and it is shown that the capacity can be approached within tenths of a dB by employing standard binary codes while avoiding the use of distribution shapers.

cs.IT↗

arXiv2010

Belief Propagation based MIMO Detection Operating on Quantized Channel Output

In multiple-antenna communications, as bandwidth and modulation order increase, system components must work with demanding tolerances. In particular, high resolution and high sampling rate analog-to-digital converters (ADCs) are often prohibitively challenging to design. Therefore ADCs for such applications should be low-resolution. This paper provides new insights into the problem of optimal signal detection based on quantized received signals for multiple-input multiple-output (MIMO) channels. It capitalizes on previous works which extensively analyzed the unquantized linear vector channel using graphical inference methods. In particular, a "loopy" belief propagation-like (BP) MIMO detection algorithm, operating on quantized data with low complexity, is proposed. In addition, we study the impact of finite receiver resolution in fading channels in the large-system limit by means of a state evolution analysis of the BP algorithm, which refers to the limit where the number of transmit and receive antennas go to infinity with a fixed ratio. Simulations show that the theoretical findings might give accurate results even with moderate number of antennas.

cs.IT↗

arXiv2010

Multiple Parameter Estimation With Quantized Channel Output

We present a general problem formulation for optimal parameter estimation based on quantized observations, with application to antenna array communication and processing (channel estimation, time-of-arrival (TOA) and direction-of-arrival (DOA) estimation). The work is of interest in the case when low resolution A/D-converters (ADCs) have to be used to enable higher sampling rate and to simplify the hardware. An Expectation-Maximization (EM) based algorithm is proposed for solving this problem in a general setting. Besides, we derive the Cramer-Rao Bound (CRB) and discuss the effects of quantization and the optimal choice of the ADC characteristic. Numerical and analytical analysis reveals that reliable estimation may still be possible even when the quantization is very coarse.

cs.IT↗