Adversarial Body Shape Search for Legged Robots
This work addresses the vulnerability of legged robots to adversarial body shape attacks, which is an incremental advancement for robotics security and diagnostics.
The authors tackled the problem of adversarial attacks on legged robots by proposing an evolutionary computation method to search for body shapes that minimize walking rewards, revealing vulnerabilities in specific robots like Walker2d and Ant-v2 being more affected by length changes, with Humanoid-v2 vulnerable to both length and thickness modifications.
We propose an evolutionary computation method for an adversarial attack on the length and thickness of parts of legged robots by deep reinforcement learning. This attack changes the robot body shape and interferes with walking-we call the attacked body as adversarial body shape. The evolutionary computation method searches adversarial body shape by minimizing the expected cumulative reward earned through walking simulation. To evaluate the effectiveness of the proposed method, we perform experiments with three-legged robots, Walker2d, Ant-v2, and Humanoid-v2 in OpenAI Gym. The experimental results reveal that Walker2d and Ant-v2 are more vulnerable to the attack on the length than the thickness of the body parts, whereas Humanoid-v2 is vulnerable to the attack on both of the length and thickness. We further identify that the adversarial body shapes break left-right symmetry or shift the center of gravity of the legged robots. Finding adversarial body shape can be used to proactively diagnose the vulnerability of legged robot walking.