Binary sequence pair based time-domain superimposed training for OFDM

This paper introduces a class of perfect binary sequence pair (PBSP), which is characterized by the impulse-equivalent periodic autocorrelation function (PACF) and the length of N=2^k. The proposed PBSP can be generated from a class of difference set pairs (DSP). Considering the desirable properties of the proposed sequence pair, orthogonal frequency division multiplexing (OFDM) transmission with time-domain superimposed training is proposed. Simulation results show that our binary sequence pair is ideal for channel estimation in terms of complexity, accuracy and efficiency.