Imitation Learning from MPC for Quadrupedal Multi-Gait Control

03/26/2021
by   Alexander Reske, et al.
0

We present a learning algorithm for training a single policy that imitates multiple gaits of a walking robot. To achieve this, we use and extend MPC-Net, which is an Imitation Learning approach guided by Model Predictive Control (MPC). The strategy of MPC-Net differs from many other approaches since its objective is to minimize the control Hamiltonian, which derives from the principle of optimality. To represent the policies, we employ a mixture-of-experts network (MEN) and observe that the performance of a policy improves if each expert of a MEN specializes in controlling exactly one mode of a hybrid system, such as a walking robot. We introduce new loss functions for single- and multi-gait policies to achieve this kind of expert selection behavior. Moreover, we benchmark our algorithm against Behavioral Cloning and the original MPC implementation on various rough terrain scenarios. We validate our approach on hardware and show that a single learned policy can replace its teacher to control multiple gaits.

READ FULL TEXT
research
09/11/2019

MPC-Net: A First Principles Guided Policy Search

We present an Imitation Learning approach for the control of dynamical s...
research
10/17/2022

Model Predictive Control via On-Policy Imitation Learning

In this paper, we leverage the rapid advances in imitation learning, a t...
research
09/18/2022

Dynamic Walking of Bipedal Robots on Uneven Stepping Stones via Adaptive-frequency MPC

This paper presents a novel Adaptive-frequency MPC framework for bipedal...
research
03/03/2020

MPC-guided Imitation Learning of Neural Network Policies for the Artificial Pancreas

Even though model predictive control (MPC) is currently the main algorit...
research
04/03/2023

Imitation Learning from Nonlinear MPC via the Exact Q-Loss and its Gauss-Newton Approximation

This work presents a novel loss function for learning nonlinear Model Pr...
research
05/30/2023

GAN-MPC: Training Model Predictive Controllers with Parameterized Cost Functions using Demonstrations from Non-identical Experts

Model predictive control (MPC) is a popular approach for trajectory opti...
research
09/26/2021

Linear Policies are Sufficient to Realize Robust Bipedal Walking on Challenging Terrains

In this work, we demonstrate robust walking in the bipedal robot Digit o...

Please sign up or login with your details

Forgot password? Click here to reset