A Blind Roll-off Factor Estimation Method for Wireless Signal Identification

In blind signal detection, exact roll-off factor estimation needs to be performed for matched filtering, but the performance of the present methods is far from satisfactory. Considering this, this paper proposes a new method for the roll-off factor estimation, and the root mean square error (RMSE)criterion is employed to measure its performance advantage compared with the previous estimation methods. To be more specific, the envelope of the impulse response of the raised cosine-roll-off filter is derived with the help of the inverse Fourier transforms, and the roll-off factor is estimated by the ratio of the minimum to the maximum of the envelope. Simulation results show that the RMSE performance of the proposed method can be improved by 78% compared with that of the existing method when the SNR is 15dB. Moreover, the RMSE performance can be found to decrease dramatically with the growth of the roll-off factor when the factor is larger than 0.6.