Xiaolei Liu

Northeast Petroleum University, China Academy of Engineering Physics

Papers

Total Citations

H-Index

About

Xiaolei Liu’s research spans the critical intersection of robotics, autonomous systems, and multi-agent security, with a focus on real-world safety and adversarial robustness. His early work established foundational methods for autonomous hazard detection, including a mobile robot system for identifying dangerous gas environments (12 citations) and a laser range finder-based approach for tracking and avoiding moving obstacles in dynamic settings (5 citations). These contributions advanced the practical deployment of robots in unpredictable, safety-critical spaces. More recently, Liu has tackled emerging security challenges in artificial intelligence, co-authoring the highly influential paper “SUB-PLAY: Adversarial Policies against Partially Observed Multi-Agent Reinforcement Learning Systems” (6 citations). This work reveals how adversarial agents can exploit partial observability in multi-agent reinforcement learning (MARL)—a vulnerability with profound implications for swarm drone control, robotic manipulation, and multi-target encirclement. By exposing these threats, Liu’s research is driving the development of more resilient, secure autonomous systems. His trajectory from sensor-based navigation to cutting-edge AI security demonstrates a sustained commitment to ensuring that intelligent machines operate safely in complex, contested environments.