论文部分内容阅读
Signals arrive out of phase at the intended receiver from collaborative beamforming(CB) nodes due to the instability in the output frequency signals of the universal software radio peripheral’s(USRP) local oscillator(LO). These nodes including the target must synchronize their oscillator frequencies for coherent signal reception. In order to do this, frequencies and phases of the signals should be estimated in software defined radio(SDR) and smoothen with nonlinear filters such as the extended Kalman filter(EKF). The process noise parameters of the NI USRP-2920 nodes will have to be calculated and used with the EKF process noise covariance matrix. These nodes are green communication hardware devices where most of the hardware units are now software defined. This article uses the direct spectrum method to obtain the phase noise values at various frequency offsets of the NI USRP-2920 in order to calculate the power spectral density of fractional frequency fluctuation. By applying the power-law noise model to this obtained value, the generated white frequency noise and random walk frequency noise values are q_1=1.93×10~(-21) and q_2=5.86×10~(-18), respectively.
Signals arrive out of phase at the intended receiver from collaborative beamforming (CB) nodes due to the instability in the output frequency signals of the universal software radio peripheral’s (USRP) local oscillator (LO). These nodes include the target must synchronize their oscillator frequencies for coherent signal reception. In order to do this, frequencies and phases of the signals should be estimated in software defined radio (SDR) and smoothen with nonlinear filters such as the extended Kalman filter (EKF). The process noise parameters of the NI USRP -2920 nodes will have to be calculated and used with the EKF process noise covariance matrix. These nodes use green software hardware devices where most of the hardware units are now software defined. This article uses the direct spectrum method to obtain the phase noise values at various frequency offsets of the NI USRP-2920 in order to calculate the power spectral density of fractional frequency fluctuation. By applying the pow er-law noise model to this obtained value, the generated white frequency noise and random walk frequency noise values are q_1 = 1.93 × 10 ~ (-21) and q_2 = 5.86 × 10 ~ (-18), respectively.