Filter-Aware Model-Predictive Control

Baris Kayalibay, Atanas Mirchev, Ahmed Agha, Patrick van der Smagt, Justin Bayer
Proceedings of The 5th Annual Learning for Dynamics and Control Conference, PMLR 211:1441-1454, 2023.

Abstract

Partially-observable problems pose a trade-off between reducing costs and gathering information. They can be solved optimally by planning in belief space, but that is often prohibitively expensive. Model-predictive control (MPC) takes the alternative approach of using a state estimator to form a belief over the state, and then plan in state space. This ignores potential future observations during planning and, as a result, cannot actively increase or preserve the certainty of its own state estimate. We find a middle-ground between planning in belief space and completely ignoring its dynamics by only reasoning about its future accuracy. Our approach, filter-aware MPC, penalises the loss of information by what we call “trackability”, the expected error of the state estimator. We show that model-based simulation allows condensing trackability into a neural network, which allows fast planning. In experiments involving visual navigation, realistic every-day environments and a two-link robot arm, we show that filter-aware MPC vastly improves regular MPC.

Cite this Paper

BibTeX
@InProceedings{pmlr-v211-kayalibay23a, title = {Filter-Aware Model-Predictive Control}, author = {Kayalibay, Baris and Mirchev, Atanas and Agha, Ahmed and Smagt, Patrick van der and Bayer, Justin}, booktitle = {Proceedings of The 5th Annual Learning for Dynamics and Control Conference}, pages = {1441--1454}, year = {2023}, editor = {Matni, Nikolai and Morari, Manfred and Pappas, George J.}, volume = {211}, series = {Proceedings of Machine Learning Research}, month = {15--16 Jun}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v211/kayalibay23a/kayalibay23a.pdf}, url = {https://proceedings.mlr.press/v211/kayalibay23a.html}, abstract = {Partially-observable problems pose a trade-off between reducing costs and gathering information. They can be solved optimally by planning in belief space, but that is often prohibitively expensive. Model-predictive control (MPC) takes the alternative approach of using a state estimator to form a belief over the state, and then plan in state space. This ignores potential future observations during planning and, as a result, cannot actively increase or preserve the certainty of its own state estimate. We find a middle-ground between planning in belief space and completely ignoring its dynamics by only reasoning about its future accuracy. Our approach, filter-aware MPC, penalises the loss of information by what we call “trackability”, the expected error of the state estimator. We show that model-based simulation allows condensing trackability into a neural network, which allows fast planning. In experiments involving visual navigation, realistic every-day environments and a two-link robot arm, we show that filter-aware MPC vastly improves regular MPC.} }
Endnote
%0 Conference Paper %T Filter-Aware Model-Predictive Control %A Baris Kayalibay %A Atanas Mirchev %A Ahmed Agha %A Patrick van der Smagt %A Justin Bayer %B Proceedings of The 5th Annual Learning for Dynamics and Control Conference %C Proceedings of Machine Learning Research %D 2023 %E Nikolai Matni %E Manfred Morari %E George J. Pappas %F pmlr-v211-kayalibay23a %I PMLR %P 1441--1454 %U https://proceedings.mlr.press/v211/kayalibay23a.html %V 211 %X Partially-observable problems pose a trade-off between reducing costs and gathering information. They can be solved optimally by planning in belief space, but that is often prohibitively expensive. Model-predictive control (MPC) takes the alternative approach of using a state estimator to form a belief over the state, and then plan in state space. This ignores potential future observations during planning and, as a result, cannot actively increase or preserve the certainty of its own state estimate. We find a middle-ground between planning in belief space and completely ignoring its dynamics by only reasoning about its future accuracy. Our approach, filter-aware MPC, penalises the loss of information by what we call “trackability”, the expected error of the state estimator. We show that model-based simulation allows condensing trackability into a neural network, which allows fast planning. In experiments involving visual navigation, realistic every-day environments and a two-link robot arm, we show that filter-aware MPC vastly improves regular MPC.
APA
Kayalibay, B., Mirchev, A., Agha, A., Smagt, P.v.d. & Bayer, J.. (2023). Filter-Aware Model-Predictive Control. Proceedings of The 5th Annual Learning for Dynamics and Control Conference, in Proceedings of Machine Learning Research 211:1441-1454 Available from https://proceedings.mlr.press/v211/kayalibay23a.html.

Related Material