Composition Learning in “Modular” Robot Systems

Abstract

Modular robot and multi-robot systems share a concept in common: composition, i.e. the study of how parts can be combined so they can be used to achieve certain objectives. Our vision is to enable robotic systems to configure and reconfigure themselves during field deployment, either autonomously or with the help of users, to adapt to emerging tasks and conditions. This goal requires us to generate compositions in real-time, while maintaining the ability to handle emergent constraints and conflicting objectives. To address these challenges, we present evolution-guided generative adversarial networks (EG-GAN) that learns to map task to compositions. Our method trains a generative model to map a task to a distribution of compositions, with training signals guided by the output of evolutionary algorithm operations. Once trained, the EG-GAN can be used to produce compositions in a near real-time fashion. We demonstrate the effectiveness of our algorithm on two distinct composition problems: 1. designing modular robots and 2. forming teams for multi-robot systems, and show that our algorithm outperforms the previous state-of-the-art algorithms in solution quality, solution diversity, and the ability to handle multiple objectives.

A separate challenge in robot composition involves the complexity introduced by inter-component connectivity, which makes the composition space high-dimensional and topologically diverse, and therefore hard to search within. We introduce Grammar-guided Latent Space Optimization (GLSO), a framework that transforms the original composition space into a low-dimensional, continuous latent space via unsupervised learning. The transformation converts the composition problem into a continuous optimization problem, where we apply sample-efficient Bayesian Optimization to search in the latent space for high-performing compositions. Our method allows us to search in the high-dimensional robot composition space more efficiently than previous state-of-the-art.