PIXEL: A Novel Detection Algorithm for Super Constellations

The demand for higher data rates in next-generation wireless networks has led to the emergence of super constellations that support large symbol sets while remaining robust to phase noise. However, efficient detection in such constellations is challenging due to the absence of low-complexity methods that operate independently of the constellation structure. Existing approaches, such as the Gaussian amplitude phase detector (GAP-D), provide accurate detection under additive Gaussian white noise (AWGN) and Gaussian phase noise (GPN), but induce highly irregular decision regions in the in-phase/quadrature (I/Q) plane, necessitating an exhaustive search over all constellation symbols that leads to prohibitive complexity in super-constellation regimes. In this letter, we propose PIXEL, a general-purpose detection algorithm that achieves constant-time O(1) complexity by pixelizing the I/Q plane into a uniform grid, with each cell linked to a small, precomputed set of candidate symbols derived from the GAP-D metric. This enables efficient, near-optimal detection without relying on structured geometries or exhaustive search. Numerical results demonstrate that PIXEL delivers high detection accuracy across diverse constellation designs, providing a flexible and practical solution for high-throughput communication systems.