Incremental Sparse GP Regression for Continuous-time Trajectory Estimation and MappingReportar como inadecuado


Incremental Sparse GP Regression for Continuous-time Trajectory Estimation and Mapping


Incremental Sparse GP Regression for Continuous-time Trajectory Estimation and Mapping - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Recent work on simultaneous trajectory estimationand mapping STEAM for mobile robots has found successby representing the trajectory as a Gaussian process. Gaussianprocesses can represent a continuous-time trajectory, elegantlyhandle asynchronous and sparse measurements, and allow therobot to query the trajectory to recover its estimated positionat any time of interest. A major drawback of this approachis that STEAM is formulated as a batch estimation problem.In this paper we provide the critical extensions necessary totransform the existing batch algorithm into an extremely efficientincremental algorithm. In particular, we are able to vastly speedup the solution time through efficient variable reordering andincremental sparse updates, which we believe will greatly increasethe practicality of Gaussian process methods for robot mappingand localization. Finally, we demonstrate the approach and itsadvantages on both synthetic and real datasets.



Computational Perception and Robotics - Computational Perception and Robotics Publications -



Autor: Yan, Xinyan - Indelman, Vadim - Boots, Byron - -

Fuente: https://smartech.gatech.edu/







Documentos relacionados