Hengrui Zhang

Yancheng Institute of Technology, Beijing Jiaotong University

Papers

Total Citations

H-Index

About

Hengrui Zhang is a rising researcher at the forefront of intelligent robotics and safe reinforcement learning. His work primarily focuses on solving two critical challenges in autonomous systems: efficient path planning for mobile robots and the development of safe, constrained reinforcement learning algorithms. In his highly cited 2022 paper, Zhang introduced a novel deep reinforcement learning approach for mobile robot path planning, employing a reflective reward design based on potential energy functions and integrating multi-agent and multi-task learning concepts—a contribution that has already garnered 22 citations for its practical impact on autonomous navigation. More recently, in 2024, Zhang has advanced the field of safety-critical AI with his work on enhancing off-policy constrained reinforcement learning. He proposed an adaptive ensemble C estimation method, directly addressing the limitations of traditional on-policy CRL algorithms by enabling more sample-efficient and safer policy learning for real-world agents. This work is pivotal for deploying RL in domains where constraint violations are costly, such as autonomous driving and robotics. Through these contributions, Zhang is establishing himself as a key innovator in bridging the gap between theoretical RL advances and their safe, practical deployment.