First, the PF is designed as an approximate resolution of the Bayesian filtering equations utilizing sequential importance sampling . For N→∞, the PF resolution recovers the true filtering density. Second, the samples in basic PF variants are generated