Wrapped Particle Filtering for Angular Data

Particle filtering is probably the most widely accepted methodology for general nonlinear applications. The performance of a particle filter critically depends on choice proposal distribution. In this paper, we propose using wrapped normal distribution as angular data, i.e. data within finite range (-π,π]. We then use same method to derive density filter, in place standard assumed Gaussian such unscented Kalman filter. numerical integrals with respect are evaluated Rogers-Szegő quadrature. Compared and similar approximate filters produce densities, show through examples that gives far better when working data. addition, demonstrate trade-off involved local sampling global (i.e. by running bank vs single filter) former yielding than latter at cost increased computational load.