Aiming at the characteristic of target tracking in microphone array networks
a new algorithm of improved Gaussian particle filter(GPF) is presented.The algorithm updates Gaussian distribution parameters of the particle with extended Kalman particle filter(EPF) rather than update all particles one by one in traditional GPF.Compared with EPF and GPF
the results of Monte-Carlo simulations demonstrate that the proposed new algorithm has higher precision and stability.
On sequential Monte Carlo sampling methods for Bayesian filtering. [J] . Arnaud Doucet,Simon Godsill,Christophe Andrieu. Statistics and Computing . 2000 (3)
Bayesian Forecasting [J] . P. J. Harrison,C. F. Stevens. Journal of the Royal Statistical Society. Series . 1976 (3)