[Paper Review] Precoding for 2x2 Doubly-Dispersive WSSUS Channels

Peter Jung|ArXiv.org|Oct 20, 2005

Optical Network Technologies12 references3 citations

TL;DR

This paper solves the optimal precoding and multiplexing problem for 2×2 doubly-dispersive WSSUS channels using a trace-class operator formulation, deriving analytic solutions via Bloch parameterization. The key result is that optimal precoders correspond to rank-one projectors onto specific states, achieving maximum channel fidelity and SINR optimality with explicit time-frequency localization properties.

ABSTRACT

Optimal link adaption to the scattering function of wide sense stationary uncorrelated scattering (WSSUS) mobile communication channels is still an unsolved problem despite its importance for next-generation system design. In multicarrier transmission such link adaption is performed by pulse shaping which in turn is equivalent to precoding with respect to the second order channel statistics. In the present framework a translation of the precoder optimization problem into an optimization problem over trace class operators is used. This problem which is also well-known in the context of quantum information theory is unsolved in general due to its non-convex nature. However in very low dimension the problem formulation reveals an additional analytic structure which again admits the solution to the optimal precoder and multiplexing scheme. Hence, in this contribution the analytic solution of the problem for the 2x2 doubly--dispersive WSSUS channel is presented.

Motivation & Objective

• To address the unsolved problem of optimal link adaptation in doubly-dispersive WSSUS channels with only second-order channel statistics available at the transmitter.
• To formulate the precoder optimization problem as a trace-class operator minimization, leveraging connections to quantum information theory.
• To derive an analytic solution for the optimal precoding and multiplexing scheme in the minimal 2×2 dimension.
• To demonstrate that channel-fidelity optimality aligns with SINR optimality in this low-dimensional setting.
• To identify explicit precoder structures that exhibit time-frequency localization, analogous to continuous Weyl-Heisenberg signaling.

Proposed method

• Formulates the precoding problem as an optimization over trace-class operators, mapping the WSSUS pulse shaping problem to a quantum-inspired fidelity maximization.
• Uses the Bloch sphere parameterization to represent density matrices and derive optimal solutions in the 2×2 case.
• Applies the Weyl-Heisenberg group representation in finite-dimensional space ℂ² to model time-frequency shifts.
• Derives optimal precoders as rank-one projectors X(opt)(n) = ½(σ₀ + σₙ) for n=1,2,3, corresponding to specific signal states.
• Establishes transmitter-side orthogonality conditions to enable direct multiplexing without additional orthogonalization.
• Verifies that the optimal precoders achieve both maximum channel fidelity and SINR optimality.

Experimental results

Research questions

• RQ1 What is the analytic form of the optimal precoder for a 2×2 doubly-dispersive WSSUS channel under second-order channel statistics?
• RQ2 How does channel-fidelity optimality relate to SINR optimality in this low-dimensional setting?
• RQ3 Can transmitter-side orthogonality be achieved without additional signal processing in the 2×2 case?
• RQ4 What is the time-frequency localization behavior of the optimal precoders?
• RQ5 How does the structure of the Weyl-Heisenberg group in ℂ² enable explicit solution derivation?

Key findings

• The optimal precoders are rank-one projectors given by X(opt)(n) = ½(σ₀ + σₙ), corresponding to specific Bloch sphere states.
• Precoder x(1) = 1/√2(1,1)T is maximally spread in time and localized in frequency, while x(3) = (0,1)T is localized in time and spread in frequency.
• The precoder x(2) = 1/√2(1,i)T corresponds to a 45° rotation in the time-frequency plane, balancing localization.
• For each optimal precoder, a corresponding multiplexing scheme exists (e.g., time-division for x(3), frequency-division for x(1)) that ensures transmitter-side orthogonality.
• The solution achieves both maximum channel fidelity and SINR optimality, confirming equivalence in this setting.
• The results demonstrate that optimal signaling in the 2×2 case exhibits structured time-frequency concentration, mirroring expectations from the continuous case.