A Caltech Library Service

ABC-LMPC: Safe Sample-Based Learning MPC for Stochastic Nonlinear Dynamical Systems with Adjustable Boundary Conditions

Thananjeyan, Brijen and Balakrishna, Ashwin and Rosolia, Ugo and Gonzalez, Joseph E. and Ames, Aaron and Goldberg, Ken (2021) ABC-LMPC: Safe Sample-Based Learning MPC for Stochastic Nonlinear Dynamical Systems with Adjustable Boundary Conditions. In: Algorithmic Foundations of Robotics XIV. Springer Proceedings in Advanced Robotics. No.17. Springer , Cham, pp. 1-17. ISBN 9783030667221.

[img] PDF - Submitted Version
See Usage Policy.

[img] PDF - Supplemental Material
See Usage Policy.


Use this Persistent URL to link to this item:


Sample-based learning model predictive control (LMPC) strategies have recently attracted attention due to their desirable theoretical properties and good empirical performance on robotic tasks. However, prior analysis of LMPC controllers for stochastic systems has mainly focused on linear systems in the iterative learning control setting. We present a novel LMPC algorithm, Adjustable Boundary Condition LMPC (ABC-LMPC), which enables rapid adaptation to novel start and goal configurations and theoretically show that the resulting controller guarantees iterative improvement in expectation for stochastic nonlinear systems. We present results with a practical instantiation of this algorithm and experimentally demonstrate that the resulting controller adapts to a variety of initial and terminal conditions on 3 stochastic continuous control tasks.

Item Type:Book Section
Related URLs:
URLURL TypeDescription ReadCube access Paper
Thananjeyan, Brijen0000-0003-1841-5071
Balakrishna, Ashwin0000-0002-3508-7850
Rosolia, Ugo0000-0002-1682-0551
Gonzalez, Joseph E.0000-0003-2921-956X
Ames, Aaron0000-0003-0848-3177
Goldberg, Ken0000-0001-6747-9499
Additional Information:© The Author(s), under exclusive license to Springer Nature Switzerland AG 2021. First Online: 09 February 2021. This research was performed at the AUTOLAB at UC Berkeley in affiliation with the Berkeley AI Research (BAIR) Lab. Authors were also supported by the Scalable Collaborative Human-Robot Learning (SCHooL) Project, a NSF National Robotics Initiative Award 1734633, and in part by donations from Google and Toyota Research Institute. Ashwin Balakrishna is supported by an NSF GRFP. This article solely reflects the opinions and conclusions of its authors and does not reflect the views of the sponsors. We thank our colleagues who provided helpful feedback and suggestions, especially Michael Danielczuk, Daniel Brown and Suraj Nair.
Funding AgencyGrant Number
Toyota Research InstituteUNSPECIFIED
NSF Graduate Research FellowshipUNSPECIFIED
Subject Keywords:Model predictive control; Control theory; Imitation learning
Series Name:Springer Proceedings in Advanced Robotics
Issue or Number:17
Record Number:CaltechAUTHORS:20210302-153300449
Persistent URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:108278
Deposited By: Tony Diaz
Deposited On:03 Mar 2021 19:25
Last Modified:16 Nov 2021 19:10

Repository Staff Only: item control page