Amos Lapidoth

Fading channels: how perfect need "perfect side information" be?

The analysis of flat-fading channels is often performed under the assumption that the additive noise is white and Gaussian, and that the receiver has precise knowledge of the realization of the fading process. These assumptions imply the optimality of Gaussian codebooks and of scaled nearest-neighbor decoding. Here we study the robustness of this communication scheme with respect to errors in the estimation of the fading process. We quantify the degradation in performance that results from such estimation errors, and demonstrate the lack of robustness of this scheme. For some situations we suggest the rule of thumb that, in order to avoid degradation, the estimation error should be negligible compared to the reciprocal of the signal-to-noise ratio (SNR).

On information rates for mismatched decoders

Reliable transmission over a discrete-time memoryless channel with a decoding metric that is not necessarily matched to the channel (mismatched decoding) is considered. It is assumed that the encoder knows both the true channel and the decoding metric. The lower bound on the highest achievable rate found by Csiszar and Korner (1981) and by Hui (1983) for DMCs, hereafter denoted C/sub LM/, is shown to bear some interesting information-theoretic meanings. The bound C/sub LM/ turns out to be the highest achievable rate in the random coding sense, namely, the random coding capacity for mismatched decoding. It is also demonstrated that the /spl epsiv/-capacity associated with mismatched decoding cannot exceed C/sub LM/. New bounds and some properties of C/sub LM/ are established and used to find relations to the generalized mutual information and to the generalized cutoff rate. The expression for C/sub LM/ is extended to a certain class of memoryless channels with continuous input and output alphabets, and is used to calculate C/sub LM/ explicitly for several examples of theoretical and practical interest. Finally, it is demonstrated that in contrast to the classical matched decoding case, here, under the mismatched decoding regime, the highest achievable rate depends on whether the performance criterion is the bit error rate or the message error probability and whether the coding strategy is deterministic or randomized. >

Mismatched decoding revisited: general alphabets, channels with memory, and the wide-band limit

The mismatch capacity of a channel is the highest rate at which reliable communication is possible over the channel with a given (possibly suboptimal) decoding rule. This quantity has been studied extensively for single-letter decoding rules over discrete memoryless channels (DMCs). Here we extend the study to memoryless channels with general alphabets and to channels with memory with possibly non-single-letter decoding rules. We also study the wide-band limit, and, in particular, the mismatch capacity per unit cost, and the achievable rates on an additive-noise spread-spectrum system with single-letter decoding and binary signaling.

On the Capacity of Free-Space Optical Intensity Channels

Upper and lower bounds are derived on the capacity of the free-space optical intensity channel. This channel has a nonnegative input (representing the transmitted optical intensity), which is corrupted by additive white Gaussian noise. To preserve the battery and for safety reasons, the input is constrained in both its average and its peak power. For a fixed ratio of the allowed average power to the allowed peak power, the difference between the upper and the lower bound tends to zero as the average power tends to infinity and their ratio tends to one as the average power tends to zero. When only an average power constraint is imposed on the input, the difference between the bounds tends to zero as the allowed average power tends to infinity, and their ratio tends to a constant as the allowed average power tends to zero.

Sending a Bivariate Gaussian Over a Gaussian MAC

We study the power-versus-distortion tradeoff for the distributed transmission of a memoryless bivariate Gaussian source over a two-to-one average-power limited Gaussian multiple-access channel. In this problem, each of two separate transmitters observes a different component of a memoryless bivariate Gaussian source. The two transmitters then describe their source component to a common receiver via an average-power constrained Gaussian multiple-access channel. From the output of the multiple-access channel, the receiver wishes to reconstruct each source component with the least possible expected squared-error distortion. Our interest is in characterizing the distortion pairs that are simultaneously achievable on the two source components. We focus on the ?equal bandwidth? case, where the source rate in source-symbols per second is equal to the channel rate in channel-uses per second. We present sufficient conditions and necessary conditions for the achievability of a distortion pair. These conditions are expressed as a function of the channel signal-to-noise ratio (SNR) and of the source correlation. In several cases, the necessary conditions and sufficient conditions are shown to agree. In particular, we show that if the channel SNR is below a certain threshold, then an uncoded transmission scheme is optimal. Moreover, we introduce a ?source-channel vector-quantizer? scheme which is asymptotically optimal as the SNR tends to infinity.

The Gaussian MAC with conferencing encoders

We derive the capacity region of the Gaussian version of Willemspsilas two-user MAC with conferencing encoders. This setting differs from the classical MAC in that, prior to each transmission block, the two transmitters can communicate with each other over noise-free bit-pipes of given capacities. The derivation requires a new technique for proving the optimality of Gaussian input distributions in certain mutual information maximizations under a Markov constraint. We also consider a Costa-type extension of the Gaussian MAC with conferencing encoders. In this extension, the channel can be described as a two-user MAC with Gaussian noise and Gaussian interference where the interference is known non-causally to the encoders but not to the decoder. We show that as in Costas setting the interference sequence can be perfectly canceled, i.e., that the capacity region without interference can be achieved.

On the Capacity of the Discrete-Time Poisson Channel

The large-inputs asymptotic capacity of a peak-power and average-power limited discrete-time Poisson channel is derived using a new firm (nonasymptotic) lower bound and an asymptotic upper bound. The upper bound is based on the dual expression for channel capacity and the notion of capacity-achieving input distributions that escape to infinity. The lower bound is based on a lower bound on the entropy of a conditionally Poisson random variable in terms of the differential entropy of its conditional mean.

Sending a Bi-Variate Gaussian Source over a Gaussian MAC

We study the power-versus-distortion tradeoff for the distributed transmission of a memoryless bivariate Gaussian source over a two-to-one average-power limited Gaussian multiple-access channel. In this problem, each of two separate transmitters observes a different component of a memoryless bivariate Gaussian source. The two transmitters then describe their source component to a common receiver via an average-power constrained Gaussian multiple-access channel. From the output of the multiple-access channel, the receiver wishes to reconstruct each source component with the least possible expected squared-error distortion. Our interest is in characterizing the distortion pairs that are simultaneously achievable on the two source components. We focus on the ¿equal bandwidth¿ case, where the source rate in source-symbols per second is equal to the channel rate in channel-uses per second. We present sufficient conditions and necessary conditions for the achievability of a distortion pair. These conditions are expressed as a function of the channel signal-to-noise ratio (SNR) and of the source correlation. In several cases, the necessary conditions and sufficient conditions are shown to agree. In particular, we show that if the channel SNR is below a certain threshold, then an uncoded transmission scheme is optimal. Moreover, we introduce a ¿source-channel vector-quantizer¿ scheme which is asymptotically optimal as the SNR tends to infinity.We consider a problem where a memoryless bi-variate Gaussian source is to be transmitted over an additive white Gaussian multiple-access channel with two transmitting terminals and one receiving terminal. The first transmitter only sees the first source component and the second transmitter only sees the second source component. We are interested in the pair of mean squared-error distortions at which the receiving terminal can reproduce each of the source components. It is demonstrated that in the symmetric case, below a certain signal-to-noise ratio (SNR) threshold, which is determined by the source correlation, uncoded communication is optimal. For SNRs above this threshold we present outer and inner bounds on the achievable distortions

Carbon Copying Onto Dirty Paper

A generalization of the problem of writing on dirty paper is considered in which one transmitter sends a common message to multiple receivers. Each receiver experiences on its link an additive interference (in addition to the additive noise), which is known noncausally to the transmitter but not to any of the receivers. Applications range from wireless multiple-antenna multicasting to robust dirty paper coding. We develop results for memoryless channels in Gaussian and binary special cases. In most cases, we observe that the availability of side information at the transmitter increases capacity relative to systems without such side information, and that the lack of side information at the receivers decreases capacity relative to systems with such side information. For the noiseless binary case, we establish the capacity when there are two receivers. When there are many receivers, we show that the transmitter side information provides a vanishingly small benefit. When the interference is large and independent across the users, we show that time sharing is optimal. For the Gaussian case, we present a coding scheme and establish its optimality in the high signal-to-interference-plus-noise limit when there are two receivers. When the interference power is large and independent across all the receivers, we show that time-sharing is again optimal. Connections to the problem of robust dirty paper coding are also discussed

Increased capacity per unit-cost by oversampling

It is demonstrated that doubling the sampling rate recovers some of the loss in capacity incurred on the bandlimited Gaussian channel with a one-bit output quantizer.