[Paper Review] Optimal Power Distribution and Minimum Expected Distortion in Gaussian Layered Broadcast Coding with Successive Refinement

Chris T. K. Ng, Denız Gündüz+2 more|arXiv (Cornell University)|May 22, 2007

Advanced MIMO Systems Optimization18 references6 citations

TL;DR

This paper proposes an optimal power allocation strategy for Gaussian layered broadcast coding over slowly fading channels without channel state information at the transmitter (CSI). By successively refining source layers and allocating power based on fading distribution, it minimizes expected distortion; for two layers, power is first assigned to the higher layer up to a fading-dependent ceiling before allocating residual power to the lower layer, with convergence to capacity-optimizing power distribution as bandwidth ratio b→0.

ABSTRACT

A transmitter without channel state information (CSI) wishes to send a delay-limited Gaussian source over a slowly fading channel. The source is coded in superimposed layers, with each layer successively refining the description in the previous one. The receiver decodes the layers that are supported by the channel realization and reconstructs the source up to a distortion. The expected distortion is minimized by optimally allocating the transmit power among the source layers. For two source layers, the allocation is optimal when power is first assigned to the higher layer up to a power ceiling that depends only on the channel fading distribution; all remaining power, if any, is allocated to the lower layer. In the limit of a continuum of infinite layers, the optimal power distribution is given by the solution to a set of linear differential equations in terms of the density of the fading distribution. As the bandwidth ratio b (channel uses per source symbol) tends to zero, the power distribution that minimizes expected distortion converges to the one that maximizes expected capacity. While expected distortion can be improved by acquiring CSI at the transmitter (CSIT) or by increasing diversity from the realization of independent fading paths, at high SNR the performance benefit from diversity exceeds that from CSIT, especially

Motivation & Objective

• To minimize expected distortion in delay-limited transmission of a Gaussian source over a slowly fading channel with no channel state information at the transmitter (CSI).
• To determine the optimal power distribution across superimposed source layers in a layered broadcast coding scheme with successive refinement.
• To analyze the impact of fading distribution and bandwidth ratio on power allocation and distortion performance.
• To compare the performance gains from diversity (multiple fading paths) versus channel state information at the transmitter (CSIT) at high SNR.
• To derive the asymptotic behavior of the optimal power allocation as the number of layers approaches infinity.

Proposed method

• Formulates a layered Gaussian broadcast model where each layer refines the previous one, with the receiver decoding only layers supported by the current channel realization.
• Derives optimal power allocation for two layers by first assigning power to the higher layer up to a ceiling determined solely by the fading distribution's cumulative distribution function.
• Extends the analysis to the limit of infinitely many layers, modeling the optimal power distribution as the solution to a system of linear differential equations derived from the fading distribution's density.
• Analyzes the asymptotic behavior as the bandwidth ratio b (channel uses per source symbol) tends to zero, showing convergence of the distortion-minimizing power allocation to the capacity-maximizing distribution.
• Compares performance with and without CSIT and with diversity gain from independent fading paths, using high-SNR asymptotics to evaluate relative gains.

Experimental results

Research questions

• RQ1 How should transmit power be optimally allocated across layered Gaussian source descriptions when the transmitter lacks channel state information?
• RQ2 What is the structure of the optimal power allocation for two-layer layered broadcast coding under a slowly fading channel?
• RQ3 How does the optimal power distribution behave in the limit of an infinite number of source layers?
• RQ4 What is the relationship between the distortion-minimizing power allocation and the capacity-maximizing allocation as the bandwidth ratio b approaches zero?
• RQ5 How do diversity gains from independent fading paths compare to the benefits of acquiring CSI at the transmitter in terms of distortion reduction at high SNR?

Key findings

• For two-layer coding, optimal power allocation assigns power to the higher layer up to a ceiling determined exclusively by the fading distribution’s cumulative distribution function, with any remaining power allocated to the lower layer.
• In the infinite-layer limit, the optimal power distribution is determined by solving a system of linear differential equations derived from the fading distribution’s probability density function.
• As the bandwidth ratio b approaches zero, the distortion-minimizing power allocation converges to the power distribution that maximizes expected channel capacity.
• At high SNR, the performance gain from diversity via independent fading paths exceeds that from acquiring CSI at the transmitter, especially in terms of distortion reduction.
• The optimal power allocation strategy ensures minimum expected distortion without requiring CSI at the transmitter, relying instead on statistical knowledge of the fading distribution.