Title :
Fast sequential Monte Carlo PHD smoothing
Author :
Nagappa, Sharad ; Clark, Daniel E.
Author_Institution :
Sch. of EPS, Heriot Watt Univ., Edinburgh, UK
Abstract :
This paper proposes a means to achieve tractable particle PHD smoothing through the use of an augmented state space label which tracks the evolution of particles over time. The use of the label reduces the forward-backward particle smoother from quadratic to linear complexity in the number of targets allowing smoothing to be carried out on a large number of targets as well as in the presence of moderate and high levels of clutter.
Keywords :
Monte Carlo methods; particle filtering (numerical methods); probability; smoothing methods; PHD filter; augmented state space label; forward-backward particle smoother; linear complexity; probability hypothesis density; quadratic complexity; sequential Monte Carlo PHD smoothing; tractable particle PHD smoothing; Clutter; Complexity theory; Equations; Labeling; Mathematical model; Smoothing methods; Target tracking; Finite Set Statistics; PHD filters; forward-backward smoothing;
Conference_Titel :
Information Fusion (FUSION), 2011 Proceedings of the 14th International Conference on
Conference_Location :
Chicago, IL
Print_ISBN :
978-1-4577-0267-9
