A maximum likelihood estimator (MLE) is developed for determining the trajectory of a constant velocity target from data of incertain origin. The result is a nonlinear iterative algorithm that adaptively weights the data to make a probabilistic assignment of measurement to source. The Cramer-Rao bound (CRB) is derived and reveals the degradation expected in system performances due to the nonassociated nature of the data. The assumption of accurate knowledge of the noise variance is relaxed and an estimator of this parameter is developed. The resulting algorithm initially utilizes a broadly distributed weighting function which, as the estimate of the variance improves, 'tightens up' around the collection of measurements associated with the target. This adaptive windowing is experimentally shown to be robust in the sense that the state estimate used at each stage of the iteration process need not be refined to a high degree of accuracy to provide acceptable estimator performance.
Trajectory and parameter estimation with measurements of uncertain origin
Trajektorien- und Parameterschaetzung unter Verwendung von Messungen mit unbekanntem Nullpunkt
Eighteenth ; 170-178
1984
9 Seiten, 9 Bilder, 12 Quellen
Conference paper
English
Origin destination estimation based on vehicle trajectory data
European Patent Office | 2017
|ORIGIN DESTINATION ESTIMATION BASED ON VEHICLE TRAJECTORY DATA
European Patent Office | 2016
|Time-varying parameter estimation with application to trajectory tracking
Online Contents | 2007
|Time-varying parameter estimation with application to trajectory tracking
Emerald Group Publishing | 2007
|