Nonlinear filtering for speaker tracking in noisy and reverberant environmentsParticle filtering for non-stationary speech modeling and enhancementSequential Monte Carlo fusion of sound and vision for speaker tracking