Farzad Etemadi

Rate and Power Allocation for Layered Transmission With Superposition Coding

We consider layered transmission of a Gaussian source over a quasi-static fading channel. A broadcast strategy is used in which multiple layers of source data are superimposed, and each layer is allocated a different power and transmission rate. For a multiple-antenna system where either the transmitter or the receiver has a single antenna, we propose a low-complexity algorithm for minimizing the expected end-to-end distortion of the received signal. Numerical results for a Rayleigh fading channel are presented, and the performance gain over a time-sharing strategy is quantified. It is numerically shown that for a wide range of operating conditions, equal rate allocation is as good as unequal rate allocation.

Transmission Over Slowly Fading Channels Using Unreliable Quantized Feedback

We study the problem of maximizing the expected rate over a slowly fading channel with quantized channel state information at the transmitter (CSIT). This problem has been recently studied in the literature assuming a noiseless feedback link. In this work, we consider a more realistic model, where the feedback link suffers from fading, as well as the limited power allocated to the feedback signals. Our scheme considers a finite-state model to capture the fading in the feedback link. We solve the rate maximization problem with different power control strategies at the transmitter. A channel optimized sealer quantizer (COSQ) is designed to incorporate feedback in our transmission scheme. Unlike the conventional COSQs where the objective is to reconstruct the source, our proposed quantizer is designed to optimize the expected rate of the forward link. For a high quality feedback channel, the proposed system performs close to the noiseless feedback case, while its performance converges to the no-feedback scenario as the feedback channel quality degrades

Optimal Layered Transmission Over Quasi-Static Fading Channels

We consider layered transmission of a successively refinable complex Gaussian source over a quasi-static fading channel. For a given number of source coding layers, we propose an efficient algorithm to calculate the optimal rate assignment for each layer, as well as the optimal size of each layer. The optimality of the algorithm is proved and numerical results for a multiple antenna Rayleigh fading channel are presented. It is numerically shown that a small number of layers is usually sufficient to achieve most of the layering gain

Throughput maximization over slowly fading channels using quantized and erroneous feedback

We design a conceptual transmission scheme that adjusts rate and power of data codewords to send them over a slowly fading channel, when quantized and possibly erroneous channel state information (CSI) is available at the transmitter. The goal is to maximize the data throughput or the expected data rate using a multi-layer superposition coding technique and temporal power control at the transmitter. The main challenge here is to design a CSI quantizer structure for a noisy feedback link. This structure resembles conventional joint source and channel coding schemes, however, with a newly introduced quasi-gray bit-mapping. Our results show that with proper CSI quantizer design, even erroneous feedback can provide performance gains. Also, with an unreliable feedback link, superposition coding provides significant gains when feedback channel is poorly conditioned and channel uncertainty at the transmitter is severe, whereas power control is more effective with more reliable feedback.

Distortion-optimal transmission of progressive images over channels with random bit errors and packet erasures

We present a statistical optimization framework for solving the end-to-end problem of progressive transmission of images over noisy channels. We consider the impacts of transmission bit errors as well as packet erasures. To cope with the impact of random bit errors, we formulate an optimization problem aimed at minimizing the end-to-end expected distortion of a reconstructed image subject to rate and efficiency constraints. In order to eliminate the impact of packet erasures, we propose utilizing an algorithm that is capable of statistically guaranteeing the delivery of a packet set associated with the progressive bitstream of an image source. Using receiver feedback, our framework is capable of effectively coping with the channel loss effects characterized by the Gilbert-Elliott model.

Optimal Rate and Power Allocation for Layered Transmission with Superposition Coding

We consider layered transmission of a Gaussian source over a quasi-static fading channel. A broadcast strategy is used in which multiple layers of source data are superimposed, and each layer is allocated a different power and transmission rate. For a multiple-antenna system where either the transmitter or the receiver has a single antenna, we propose a low-complexity algorithm for minimizing the expected end-to-end distortion of the received signal. Numerical results for a Rayleigh fading channel are presented, and the performance gain over a time-sharing strategy is quantified. It is numerically shown that for a wide range of operating conditions, equal rate allocation is as good as unequal rate allocation.

Outage Behavior of Slow Fading Channels With Power Control Using Partial and Erroneous CSIT

This work investigates high-SNR outage behavior of slow fading channels with quantized feedback and partial power control, when feedback indices are error-prone. We propose a quantizer structure with continuous Voronoi regions and derive its optimality conditions using probability analysis. Our design involves construction of a novel bit-mapping scheme. The optimized power control codebook resembles channel optimized scalar quantizers (COSQs). Also, the diversity gain of the communication system under study is characterized with both nonzero and vanishing feedback error probabilities.

An Efficient Progressive Bitstream Transmission System for Hybrid Channels With Memory

We consider progressive transmission over a hybrid channel introducing bit errors and packet erasures. The existing solutions are analyzed and extended to the case of a channel that exhibits memory on both bit errors and packet erasures. We then propose a simple, low-complexity coding scheme that transforms the hybrid channel into a channel with a single impairment for which various optimization techniques exist. Both rate-based and distortion-based optimization problems are investigated. It is shown that our proposed solution has lower channel coding and rate-distortion optimization complexities compared to the known solutions. Simulation results for channels with and without memory show the effectiveness of our proposed solution over a wide range of operating conditions. Numerical results also indicate that the rate-based solution of our proposed algorithm is very close to the corresponding distortion-based solution

A linear-complexity distortion optimal scheme for the transmission of packetized progressive bitstreams

We propose a novel distortion minimization technique for the transmission of a packetized progressive bitstream. The optimality of our proposed algorithm is analytically proved for a class of sources satisfying a stated condition. It is shown that Gauss-Markov sources belong to the latter class for which the algorithm is optimal. We show that our proposed optimization technique is robust and has a linear complexity in the transmission rate. Simulation results show the effectiveness of our proposed algorithm.

A Unified Framework for Layered Transmission Over Fading and Packet Erasure Channels

We consider layered transmission of a successively refinable source over a quasi-static fading channel. We establish a duality relationship between this problem and that of packet transmission over erasure channels and use it to share solution techniques in both domains. For a Gaussian source and the fading channel, a low-complexity, optimal algorithm is proposed, and it is shown that the corresponding dual for packet erasure channels has a linear complexity as opposed to the quadratic complexity of the best known optimal algorithms in the literature. For non-Gaussian sources, the optimal rate allocation problem for fading channels is solved using the dual solution for erasure channels. It is also shown that a single-layer system is optimal for fading channels if the goal is to maximize the rate. Numerical results for multiple antenna Rayleigh fading channels are presented for Gaussian sources and practical image coders. It is shown that a few number of layers significantly improves the performance. Finally, we numerically show that for practical operating conditions, optimizing the asymptotic measure of distortion exponent is not enough when there are more than one transmit or receive antennas.