Adversarial joint attacks on legged robots

TL;DR

This paper investigates the vulnerability of legged robots to adversarial torque perturbations at joints, demonstrating black-box attack methods that can induce instability, with implications for robot safety and robustness.

Contribution

It introduces black-box adversarial attack techniques for legged robots' actuators, applicable across different architectures, and evaluates their effectiveness on simulated robots.

Findings

Differential evolution efficiently finds strong adversarial perturbations.

Quadruped robot Ant-v2 is vulnerable to joint attacks.

Bipedal robot Humanoid-v2 shows robustness to perturbations.

Abstract

We address adversarial attacks on the actuators at the joints of legged robots trained by deep reinforcement learning. The vulnerability to the joint attacks can significantly impact the safety and robustness of legged robots. In this study, we demonstrate that the adversarial perturbations to the torque control signals of the actuators can significantly reduce the rewards and cause walking instability in robots. To find the adversarial torque perturbations, we develop black-box adversarial attacks, where, the adversary cannot access the neural networks trained by deep reinforcement learning. The black box attack can be applied to legged robots regardless of the architecture and algorithms of deep reinforcement learning. We employ three search methods for the black-box adversarial attacks: random search, differential evolution, and numerical gradient descent methods. In experiments with the quadruped robot Ant-v2 and the bipedal robot Humanoid-v2, in OpenAI Gym environments, we find that differential evolution can efficiently find the strongest torque perturbations among the three methods. In addition, we realize that the quadruped robot Ant-v2 is vulnerable to the adversarial perturbations, whereas the bipedal robot Humanoid-v2 is robust to the perturbations. Consequently, the joint attacks can be used for proactive diagnosis of robot walking instability.