PIMbot: A Self-Adaptive Attack Framework for Adversarial Manipulation of Multi-Robot Reinforcement Learning

2026-05-21

AI Summary

This work addresses the vulnerability of multi-robot reinforcement learning systems in social dilemmas to communication interference and adversarial attacks, which can lead to catastrophic cooperation collapse. To this end, the authors propose PIMbot, a novel framework that introduces, for the first time, an adaptive multi-objective controller into multi-agent social dilemma scenarios. PIMbot leverages dual manipulation levers—reward-channel incentive manipulation and agent-action policy manipulation—to adaptively perturb cooperative behaviors online. By integrating a tailored reward function with multi-objective optimization, the approach is validated through extensive experiments in Gazebo simulations and real-world deployment on NVIDIA Jetson Orin Nano embedded platforms. The results successfully expose critical security vulnerabilities in existing coordination mechanisms while demonstrating strong practicality and cross-platform generalization capability.

Abstract

Recent research has demonstrated the potential of reinforcement learning in effective multi-robot collaboration, particularly in social dilemmas where robots face a trade-off between self-interest and collective benefits. However, environmental factors such as miscommunication and adversarial robots can impact cooperation, making it crucial to explore how multi-robot communication can be manipulated to achieve different outcomes. This paper presents PIMbot, a framework that manipulates outcomes via two complementary levers: (i) incentive manipulation of the reward channel and (ii) policy manipulation of an agent's own actions. An adaptive multi-objective controller balances these levers in an online manner. Our work introduces a novel approach to manipulation in recent multi-agent RL social dilemmas that utilize a unique reward function for incentivization. By utilizing our proposed PIMbot mechanisms, a robot is able to manipulate the social dilemma environment effectively. Comprehensive experimental results demonstrate the effectiveness of our proposed methods in the Gazebo-simulated multi-robot environment. Moreover, a real embedded device case study on NVIDIA Jetson Orin Nano quantifies system cost and validates PIMbot's effectiveness on realistic autonomous embedded systems scenarios beyond simulation. Together, these results position PIMbot as a rigorous stress-test tool exposing critical vulnerabilities in multi-robot cooperative tasks.

Problem

• multi-robot reinforcement learning
• adversarial manipulation
• social dilemmas
• cooperative vulnerability
• reward manipulation

Innovation

• adversarial manipulation
• multi-robot reinforcement learning
• incentive manipulation
• policy manipulation
• self-adaptive attack framework