Peak-to-average power ratio reduction in OFDM: optimal and suboptimal combination of erasure pattern selection and probabilistic methods

A procedure for improving peak-to-average power ratio (PAPR) performance of probabilistic methods in OFDM is proposed. The procedure incorporates a method that introduces redundancy into the data via frame expansion. The mapping is followed by selecting the best pattern for erasing some samples before transmission. This erasure pattern selection (EPS) method can be effectively combined with existing probabilistic methods to produce a significant improvement in performance. Optimally combining the methods is computationally demanding. Suboptimal procedures proposed for complexity reduction in PTS are adapted for combining EPS with PTS. Four methods are proposed and compared through complexity analysis and simulations, and the trade-off in complexity and PAPR improvement is presented.