Lattice Particle Filters
This addresses the stochastic instability in particle filters for state-space models, offering a more reliable inference method for applications like tracking and motion estimation.
The paper tackles the performance variability of particle filters by introducing lattice particle filters, which deterministically place particles using Quasi-Monte Carlo rules, resulting in a synthetic 2D tracking performance equivalent to a conventional filter with 10-60% more particles.
A standard approach to approximate inference in state-space models isto apply a particle filter, e.g., the Condensation Algorithm.However, the performance of particle filters often varies significantlydue to their stochastic nature.We present a class of algorithms, called lattice particle filters, thatcircumvent this difficulty by placing the particles deterministicallyaccording to a Quasi-Monte Carlo integration rule.We describe a practical realization of this idea, discuss itstheoretical properties, and its efficiency.Experimental results with a synthetic 2D tracking problem show that thelattice particle filter is equivalent to a conventional particle filterthat has between 10 and 60% more particles, depending ontheir "sparsity" in the state-space.We also present results on inferring 3D human motion frommoving light displays.