Participan

Bit-interleaved coded modulation (BICM) is nowadays the most popular coded modulation (CM) scheme for fading and non-fading channels. BICM has been adopted in commercial systems such as wireless and wired broadband access networks, 3G and 4G telephony, ultrawideband transceivers, and digital video broadcasting, imposing itself as the de facto standard for current wireless telecommunications systems. Moreover, BICM will be the basis for future communication standards, and therefore, its understanding becomes crucial.

At first glance, BICM appears to be a very simple “out-of-the-box” CM scheme. However, a careful analysis reveals intriguing properties and shows many optimization opportunities. In this tutorial, we will provide a simple and comprehensive overview of BICM, and we will lead the audience through the concepts underlying BICM systems. We will explain BICM principles, compare it to the competing CM schemes (i.e., trellis-coded and multilevel coded modulation), and also explain why BICM became a de facto standard in wireless industry.

Bit-interleaved coded modulation (BICM) is nowadays the most popular coded modulation (CM) scheme for fading and non-fading channels. BICM has been adopted in commercial systems such as wireless and wired broadband access networks, 3G and 4G telephony, ultrawideband transceivers, and digital video broadcasting, imposing itself as the de facto standard for current wireless telecommunications systems. Moreover, BICM will be the basis for future communication standards, and therefore, its understanding becomes crucial.

At first glance, BICM appears to be a very simple “out-of-the-box” CM scheme. However, a careful analysis reveals intriguing properties and shows many optimization opportunities. In this tutorial, we will provide a simple and comprehensive overview of BICM, and we will lead the audience through the concepts underlying BICM systems. We will explain BICM principles, compare it to the competing CM schemes (i.e., trellis-coded and multilevel coded modulation), and also explain why BICM became a de facto standard in wireless industry.

In a number of applications, including wireless communications, it is important to estimate the probability density function of a given random source and then transmit this estimate to another point in space. The combined problem is referred to as universal density estimation and coding.This talk presents novel results on the general problem of universal density estimation under an operational data-rate constraint. We present a coding theorem that stipulates necessary and sufficient conditions to learn and transmit a memoryless source distribution with arbitrary precision (in total variations), under an asymptotic zero-rate regime, in bits per sample. In the process, we propose a concrete coding scheme to achieve this learning objective, adopting the Skeleton estimate developed by Y. Yatracos.

Attaining high energy efficiency is a key condition that wireless communications devices for wireless sensor networks must satisfy in order for the technology to prosper into large-scale, autonomous networks. In this talk, we present how physical-layer parameters like signal to noise ratio and modulation size can be optimized in order to reduce the total energy consumption per transmitted bit. For this, we have developed a model that considers the energy consumed by electromagnetic radiation, the energy used by transceiver electronics and the energy cost of feedback frames that request retransmission. We will show how the model can be used for optimizing the SNR and the constellation size for single-antenna communications over AWGN, Rayleigh and Nakagami-m fading channels and multi-antenna communications using singular value decomposition.

We consider the behavior of the BICM Receiver operating with the mismatched reliability metrics (L-values). To correct the mismatch - that may occur due to many reasons, we aim at linear correction (multiplication by a linear factor) of the L-values. Our objective is to find the correction factor that minimizes the probability of errors made by a maximum likelihood decoder that uses the corrected L-values. To this end, we use the so-called saddlepoint approximation of the pairwise error probability that makes our analysis independent of the form of the distribution of the L-values, and conclude that the correction factors should be equal to the twice of the saddlepoint of the cumulant generating function of the L-values. We provide a simple numerical example of transmission in the presence of interference where we demonstrate a notable improvement attainable with the proposed method.

This talk will present recent results on the statistical characterization of wireless fading channels. The main result is the derivation of closed-form expressions for the second-order statistics of the spectral power gain of wide-band microwave indoor channels. This results is valid within a framework general enough to be compatible with several popular channel models, such as the Saleh-Valenzuela channel model and those proposed by the IEEE 802.15.3a and IEEE 802.15.4a task groups. These models have been applied in the development of the well known WiFi technology, and are at the basis of the emerging standards for ultra-wide-band systems and body-area networks. As in all these models, our channel description is based upon clusters and rays with (possibly mixed-) Poisson arrivals and random amplitudes. Our results consist of closed-form expressions for the second-order statistics of the channel power frequency response. These expressions reveal that, perhaps surprisingly, the auto-covariance of the spectral power gain between at any two frequencies tends to zero as the difference between these frequencies tends to infinity if and only if the cluster arrival rate goes to infinity. These results allow one to obtain closed-form expressions for the variance and second-order moment of the channel power within any given interval of frequencies. From there, it possible to approximate the channel spectral diversity as an explicit function of model parameters and bandwidth.

The so-called bit-interleaved coded modulation (BICM) capacity represents an achievable rate for BICM systems and is highly dependent on the binary labeling of the constellation. A Gray code was conjectured to be the binary labeling that maximizes the BICM capacity, however, recent results have shown that in the low signal-to-noise ratio (SNR) regime Gray codes are not optimal for BICM. In this talk, we study the behavior of the BICM capacity in the high SNR regime. We develop upper and lower bounds for the BICM capacity which clearly suggest that for high SNR and an arbitrary constellation, a Gray code—if it exists—is the capacity-maximizing binary labeling for the BICM capacity. These results also indicate that a maximization of the BICM capacity in the high SNR regime is equivalent to a minimization of the average uncoded bit-error rate. The conjecture of the optimality of Gray codes is also supported by numerical examples.