Modified blind selected mapping for OFDM/single-carrier signal transmission

Blind selected mapping (blind SLM) is an effective peak-to-average power ratio (PAPR) reduction technique which does not require side information sharing. In the blind SLM, the receiver employs phase rotation sequence estimation, which can be carried out using maximum-likelihood (ML) estimation or 2-step sequence estimation using Viterbi algorithm. ML estimation typically requires much higher computational complexity. Recently, we showed that the use of codebook generated from a 2-level phase rotation set {0°, 135°} and the ML phase rotation sequence estimation based on the fourth-power QAM constellation requires much less complexity compared to the conventional blind SLM with 3-level phase rotation set {0°, 120°, 240°} and ML estimation based on original QAM constellation. However, for a large number of phase rotation sequences, the computational complexity still remains high. In this paper, in order to further reduce the computational complexity, we propose a combined use of a 2-level phase rotation set {0°, 135°} and the 2-step sequence estimation. The use of 2-level phase rotation set significantly reduces the number of branches and states in the Viterbi algorithm and hence, leads to complexity reduction. Simulation results confirm that the blind SLM using the 2-level phase rotation set and the 2-step sequence estimation has less computational complexity while achieving similar BER to the ML sequence estimation.