Residual Physics Learning and System Identification for Sim-to-real Transfer of Policies on Buoyancy Assisted Legged Robots

The light and soft characteristics of Buoyancy Assisted Lightweight Legged Unit (BALLU) robots have a great potential to provide intrinsically safe interactions in environments involving humans, unlike many heavy and rigid robots. However, their unique and sensitive dynamics impose challenges to obtaining robust control policies in the real world. In this work, we demonstrate robust sim-to-real transfer of control policies on the BALLU robots via system identification and our novel residual physics learning method, Environment Mimic (EnvMimic). First, we model the nonlinear dynamics of the actuators by collecting hardware data and optimizing the simulation parameters. Rather than relying on standard supervised learning formulations, we utilize deep reinforcement learning to train an external force policy to match real-world trajectories, which enables us to model residual physics with greater fidelity. We analyze the improved simulation fidelity by comparing the simulation trajectories against the real-world ones. We finally demonstrate that the improved simulator allows us to learn better walking and turning policies that can be successfully deployed on the hardware of BALLU.

I. Introduction

Buoyancy-assisted or balloon-based robots [1]–[3] have great potential to offer fundamental safety in human en-vironments. Traditional mobile robots while being able to execute a variety of tasks, tend to be rigid and heavy and may cause serious damage to their surroundings or themselves in case of control or perception errors. On the other hand, buoyancy-assisted robots (BARs) [4] are typically designed to be lightweight, compact, and intrinsically safe. Therefore, they can be used for various applications that require close human-robot interaction, such as education, entertainment, and healthcare. For instance, Chae et al. [1] present Buoy-ancy Assisted Lightweight Legged Unit (BALLU), which is a balloon-based robot with two legs (Fig. 1), and showcase that it can be deployed to various indoor and outdoor environments without any safety concerns.