Planning in the Continuous Domain: a Generalized Belief Space Approach for Autonomous Navigation in Unknown EnvironmentsReportar como inadecuado


Planning in the Continuous Domain: a Generalized Belief Space Approach for Autonomous Navigation in Unknown Environments


Planning in the Continuous Domain: a Generalized Belief Space Approach for Autonomous Navigation in Unknown Environments - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

We investigate the problem of planning under uncertainty, with applicationto mobile robotics. We propose a probabilistic framework inwhich the robot bases its decisions on the generalized belief, which is a probabilistic description of its own state and of external variables of interest.The approach naturally leads to a dual-layer architecture: aninner estimation layer, which performs inference to predict the outcomeof possible decisions, and an outer decisional layer which is in charge ofdeciding the best action to undertake. Decision making is entrusted toa Model Predictive Control MPC scheme. The formulation is valid forgeneral cost functions and does not discretize the state or control space,enabling planning in continuous domain. Moreover, it allows to relax theassumption of maximum likelihood observations: predicted measurementsare treated as random variables, and binary random variables are used to model the event that a measurement is actually taken by the robot. Wesuccessfully apply our approach to the problem of uncertainty-constrainedexploration, in which the robot has to perform tasks in an unknown environment, while maintaining localization uncertainty within given bounds.We present an extensive numerical analysis of the proposed approach andcompare it against related work. In practice, our planning approach producessmooth and natural trajectories and is able to impose soft upperbounds on the uncertainty. Finally, we exploit the results of this analysis to identify current limitations and show that the proposed framework can accommodate several desirable extensions.



Computational Perception and Robotics - Computational Perception and Robotics Publications -



Autor: Indelman, Vadim - Carlone, Luca - Dellaert, Frank - -

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



DESCARGAR PDF




Documentos relacionados