Multiple access interference suppression for CDMA systems via -minimization.

Area of Science:

• Signal Processing
• Telecommunications Engineering
• Optimization Theory

Background:

• Code-Division Multiple Access (CDMA) systems face challenges with multiple access interference (MAI).
• Conventional minimum output energy (MOE) detectors perform poorly with signature waveform distortions from multipath or timing issues.
• Blind multiuser detection aims to detect signals without prior knowledge of other users' information.

Purpose of the Study:

• To propose a novel robust linear programming (ROLP) algorithm for blind multiuser detection in CDMA systems.
• To enhance detector performance under signature waveform mismatch and non-Gaussian signal conditions.
• To develop a computationally efficient algorithm compared to existing robust methods.

Main Methods:

• The ROLP algorithm minimizes the L1-norm of the output signal to leverage non-Gaussian signal properties.
• It incorporates robustness against signature mismatch by constraining the magnitude response within a novel rhombus-shaped uncertainty set.
• The optimization problem is reformulated as a linear program for efficient solving.

Main Results:

• The proposed ROLP algorithm demonstrates superior performance compared to existing robust detectors in simulations.
• It shows robustness against signature waveform distortions, outperforming MOE-based techniques.
• The ROLP's computational complexity is lower than robust methods requiring second-order cone programming.

Conclusions:

• The ROLP algorithm offers an effective and computationally efficient solution for blind multiuser detection in CDMA.
• It provides robust performance in the presence of signal distortions and non-Gaussianity.
• The novel rhombus uncertainty set contributes to improved detection accuracy, approaching optimal performance bounds.