MPC-Net: A First Principles Guided Policy Search
OPEN ACCESS
Loading...
Author / Producer
Date
2020-04
Publication Type
Journal Article
ETH Bibliography
yes
Citations
Altmetric
OPEN ACCESS
Data
Rights / License
Abstract
We present an Imitation Learning approach for the control of dynamical systems with a known model.
Our policy search method is guided by solutions from MPC. Typical policy search methods of this kind minimize a distance metric between the guiding demonstrations and the learned policy. Our loss function, however, corresponds to the minimization of the control Hamiltonian, which derives from the principle of optimality. Therefore, our algorithm directly attempts to solve the optimality conditions with a parameterized class of control laws. Additionally, the proposed loss function explicitly encodes the constraints of the optimal control problem and we provide numerical evidence that its minimization achieves improved constraint satisfaction. We train a mixture-of-expert neural network architecture for controlling a quadrupedal robot and show that this policy structure is well suited for such multimodal systems. The learned policy can successfully stabilize different gaits on the real walking robot from less than 10 min of demonstration data.
Permanent link
Publication status
published
External links
Editor
Book title
Journal / series
Volume
5 (2)
Pages / Article No.
2897 - 2904
Publisher
IEEE
Event
Edition / version
Methods
Software
Geographic location
Date collected
Date created
Subject
Learning from Demonstration; Legged Robots; Optimization and Optimal Control
Organisational unit
09570 - Hutter, Marco / Hutter, Marco
Notes
Funding
166232 - Data-driven control approaches for advanced legged locomotion (SNF)
188596 - Perceptive Dynamic Locomotion on Rough Terrain (SNF)
780883 - subTerranean Haptic INvestiGator (EC)
188596 - Perceptive Dynamic Locomotion on Rough Terrain (SNF)
780883 - subTerranean Haptic INvestiGator (EC)