RMPflow: A Computational Graph for Automatic Motion Policy Generation

TL;DR

RMPflow is a new algorithm that combines local motion policies into a global, geometrically consistent policy for robotic motion planning, improving efficiency and handling complex tasks.

Contribution

It introduces RMPflow, a novel method for synthesizing global motion policies from local Riemannian Motion Policies with geometric consistency and stability guarantees.

Findings

Enables stable combination of local policies into a global policy.

Improves planning efficiency in high-DOF manipulation tasks.

Simplifies complex motion planning problems through geometric considerations.

Abstract

We develop a novel policy synthesis algorithm, RMPflow, based on geometrically consistent transformations of Riemannian Motion Policies (RMPs). RMPs are a class of reactive motion policies designed to parameterize non-Euclidean behaviors as dynamical systems in intrinsically nonlinear task spaces. Given a set of RMPs designed for individual tasks, RMPflow can consistently combine these local policies to generate an expressive global policy, while simultaneously exploiting sparse structure for computational efficiency. We study the geometric properties of RMPflow and provide sufficient conditions for stability. Finally, we experimentally demonstrate that accounting for the geometry of task policies can simplify classically difficult problems, such as planning through clutter on high-DOF manipulation systems.