Low Complexity Detection Scheme for NOFDM Systems Based on ML Detection over Hyperspheres

Nonorthogonal Frequency Division Multiplexing (NOFDM) is a digital modulation technique that promises to provide extremely high spectral efficiencies. However, this modulation scheme is seldom used in practice due to the high computational complexity involved in decoding the received signal in the presence of noise. The basic aim of this paper is to reduce this decoding complexity. Here, we propose a low complexity detection algorithm which makes use of maximum likelihood (ML) decoding not over the entire signal constellation but over a proper subset of the constellation that lies on a hypersphere thereby reducing the computational complexity for decoding. Computational complexity has been evaluated for various values of transmitted power and the result has been plotted for both ML detection algorithm and the proposed algorithm at a fixed data rate. The BER performance for both the algorithms has also been compared at a fixed data rate and the result has been plotted. The results show that the proposed algorithm is far superior to ML detection algorithm in terms of computational complexity.