Learning Model Preconditions for Planning with Multiple Models

Alex Licari LaGrassa, Oliver Kroemer
Proceedings of the 5th Conference on Robot Learning, PMLR 164:491-500, 2022.

Abstract

Different models can provide differing levels of fidelity when a robot is planning. Analytical models are often fast to evaluate but only work in limited ranges of conditions. Meanwhile, physics simulators are effective at modeling complex interactions between objects but are typically more computationally expensive. Learning when to switch between the various models can greatly improve the speed of planning and task success reliability. In this work, we learn model deviation estimators (MDEs) to predict the error between real-world states and the states outputted by transition models. MDEs can be used to define a model precondition that describes which transitions are accurately modeled. We then propose a planner that uses the learned model preconditions to switch between various models in order to use models in conditions where they are accurate, prioritizing faster models when possible. We evaluate our method on two real-world tasks: placing a rod into a box and placing a rod into a closed drawer.

Cite this Paper

BibTeX
@InProceedings{pmlr-v164-lagrassa22a, title = {Learning Model Preconditions for Planning with Multiple Models}, author = {LaGrassa, Alex Licari and Kroemer, Oliver}, booktitle = {Proceedings of the 5th Conference on Robot Learning}, pages = {491--500}, year = {2022}, editor = {Faust, Aleksandra and Hsu, David and Neumann, Gerhard}, volume = {164}, series = {Proceedings of Machine Learning Research}, month = {08--11 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v164/lagrassa22a/lagrassa22a.pdf}, url = {https://proceedings.mlr.press/v164/lagrassa22a.html}, abstract = {Different models can provide differing levels of fidelity when a robot is planning. Analytical models are often fast to evaluate but only work in limited ranges of conditions. Meanwhile, physics simulators are effective at modeling complex interactions between objects but are typically more computationally expensive. Learning when to switch between the various models can greatly improve the speed of planning and task success reliability. In this work, we learn model deviation estimators (MDEs) to predict the error between real-world states and the states outputted by transition models. MDEs can be used to define a model precondition that describes which transitions are accurately modeled. We then propose a planner that uses the learned model preconditions to switch between various models in order to use models in conditions where they are accurate, prioritizing faster models when possible. We evaluate our method on two real-world tasks: placing a rod into a box and placing a rod into a closed drawer.} }
Endnote
%0 Conference Paper %T Learning Model Preconditions for Planning with Multiple Models %A Alex Licari LaGrassa %A Oliver Kroemer %B Proceedings of the 5th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2022 %E Aleksandra Faust %E David Hsu %E Gerhard Neumann %F pmlr-v164-lagrassa22a %I PMLR %P 491--500 %U https://proceedings.mlr.press/v164/lagrassa22a.html %V 164 %X Different models can provide differing levels of fidelity when a robot is planning. Analytical models are often fast to evaluate but only work in limited ranges of conditions. Meanwhile, physics simulators are effective at modeling complex interactions between objects but are typically more computationally expensive. Learning when to switch between the various models can greatly improve the speed of planning and task success reliability. In this work, we learn model deviation estimators (MDEs) to predict the error between real-world states and the states outputted by transition models. MDEs can be used to define a model precondition that describes which transitions are accurately modeled. We then propose a planner that uses the learned model preconditions to switch between various models in order to use models in conditions where they are accurate, prioritizing faster models when possible. We evaluate our method on two real-world tasks: placing a rod into a box and placing a rod into a closed drawer.
APA
LaGrassa, A.L. & Kroemer, O.. (2022). Learning Model Preconditions for Planning with Multiple Models. Proceedings of the 5th Conference on Robot Learning, in Proceedings of Machine Learning Research 164:491-500 Available from https://proceedings.mlr.press/v164/lagrassa22a.html.

Related Material