Algorithm for unknown SNR estimation based on sequential Monte Carlo method in cluttered environment

A Hybrid Labeled Multi-Bernoulli Filter With Amplitude For Tracking Fluctuating Targets

“…In practice, the target SNR is dependent on the aspect angle and radar wave frequency [13], [15], hence it will fluctuate. The dynamics of SNR has been simply described by random walk model or Gaussian density [17], [19]. Alternatively, we employ the NCG distribution as the SNR transition density, which is well-known for modeling stochastic volatility such as financial time series [21] and wind speed intensity [27].…”

“…Hence, one has to take into account the fluctuation of the target SNR when amplitude information is used. In [17], [18], the uncertainty of the target SNR is modeled by random walk (RW) with Gaussian step size and the SNR is estimated by sequential Monte Carlo (SMC) method. Ristic et al [19] employ the Bernoulli filter with amplitude measurement for single target tracking, where the target SNR is also described by a RW process.…”