Efficient FFT pruning algorithm for non-contiguous OFDM systems

This paper presents the study of an efficient trade-off between memory requirements and performance for the implementation of FFT pruning algorithm. FFT pruning algorithm is utilized in NC-OFDM systems to simplify the FFT algorithm complexity in presence of subcarrier sparseness. State-of-the-art implementations offer good performance with the drawback of high resources utilization, i.e. data memory for storage of the configuration matrix. In this work we introduce the partial pruning algorithm as an efficient way to implement FFT pruning, obtaining a balanced trade-off between performance and resources allocation. Cycle accurate simulation results showed that even in presence of low-medium input sparseness levels the proposed algorithm can reduce the computation time of at least a 20% factor, when compared to traditional FFT algorithms and, at the same time, decreases the memory utilization up to a 20% factor over state of the art pruning algorithms.