Robot Crash Course: Learning Soft and Stylized Falling

TL;DR

This paper introduces a reinforcement learning approach for bipedal robots to perform controlled, soft falls, minimizing damage and allowing flexible end poses, with demonstrated success in simulation and real-world tests.

Contribution

It proposes a robot-agnostic reward function and a novel sampling strategy to enable robust, controlled falling with arbitrary end poses in both simulation and real-world environments.

Findings

Robots can perform controlled, soft falls in simulation.

The method reduces physical damage during falls.

Successful real-world implementation demonstrated.

Abstract

Despite recent advances in robust locomotion, bipedal robots operating in the real world remain at risk of falling. While most research focuses on preventing such events, we instead concentrate on the phenomenon of falling itself. Specifically, we aim to reduce physical damage to the robot while providing users with control over a robot's end pose. To this end, we propose a robot agnostic reward function that balances the achievement of a desired end pose with impact minimization and the protection of critical robot parts during reinforcement learning. To make the policy robust to a broad range of initial falling conditions and to enable the specification of an arbitrary and unseen end pose at inference time, we introduce a simulation-based sampling strategy of initial and end poses. Through simulated and real-world experiments, our work demonstrates that even bipedal robots can perform controlled, soft falls.