Memory Clustering using Persistent Homology for Multimodality- and Discontinuity-Sensitive Learning of Optimal Control Warm-starts

by   Wolfgang Merkt, et al.

Shooting methods are an efficient approach to solving nonlinear optimal control problems. As they use local optimization, they exhibit favorable convergence when initialized with a good warm-start but may not converge at all if provided with a poor initial guess. Recent work has focused on providing an initial guess from a learned model trained on samples generated during an offline exploration of the problem space. However, in practice the solutions contain discontinuities introduced by system dynamics or the environment. Additionally, in many cases multiple equally suitable, i.e., multi-modal, solutions exist to solve a problem. Classic learning approaches smooth across the boundary of these discontinuities and thus generalize poorly. In this work, we apply tools from algebraic topology to extract information on the underlying structure of the solution space. In particular, we introduce a method based on persistent homology to automatically cluster the dataset of precomputed solutions to obtain different candidate initial guesses. We then train a Mixture-of-Experts within each cluster to predict initial guesses and provide a comparison with modality-agnostic learning. We demonstrate our method on a cart-pole toy problem and a quadrotor avoiding obstacles, and show that clustering samples based on inherent structure improves the warm-start quality.


page 1

page 6

page 7

page 11


Neural Optimal Control using Learned System Dynamics

We study the problem of generating control laws for systems with unknown...

Learning Agile Paths from Optimal Control

Efficient motion planning algorithms are of central importance for deplo...

Discontinuity-Sensitive Optimal Control Learning by Mixture of Experts

This paper proposes a discontinuity-sensitive approach to learn the solu...

Extension of Switch Point Algorithm to Boundary-Value Problems

In an earlier paper (, the Switch Poi...

Deep ℒ^1 Stochastic Optimal Control Policies for Planetary Soft-landing

In this paper, we introduce a novel deep learning based solution to the ...

Please sign up or login with your details

Forgot password? Click here to reset