Rao–Blackwellized Particle Filters With Out-of-Sequence Measurement Processing

摘要: This paper addresses the out-of-sequence measurement (OOSM) problem for mixed linear/nonlinear state-space models, which is a class of nonlinear models with tractable, conditionally linear substructure. We develop two novel algorithms that utilize The first algorithm effectively employs Rao–Blackwellized particle filtering framework updating OOSMs, and based on storing only subset particles their weights over an arbitrary, predefined interval. second adapts backward simulation approach to update delayed (out-of-sequence) measurements, resulting in superior tracking performance. Extensive studies show efficacy our approaches terms computation time Both yield estimation improvements when compared recent filter OOSM processing; considered examples they achieve up 10% enhancements accuracy. In some cases, proposed even deliver accuracy similar lower performance bounds. Because setup common various scenarios, developed enable different types applications.