Heming Cui
Papers
10
Total Citations
48
H-Index
5
About
Heming Cui is an emerging researcher at the intersection of autonomous robotics, efficient deep learning, and robotic systems. His work spans several interconnected domains, including air-ground robot (AGR) navigation, distributed deep neural network inference, robotic manipulation, and adversarial robustness in deep reinforcement learning. Cui's most notable contributions center on intelligent navigation for air-ground robots in challenging, occlusion-prone environments. His AGRNav, OMEGA, and HE-Nav systems collectively address the critical problem of navigating cluttered, dynamically changing spaces by combining 3D semantic scene completion, energy-efficient path planning, and state space modeling — work that has already garnered over 20 citations since 2024. Beyond navigation, his research tackles the practical deployment of machine learning on robotic IoT systems, proposing distributed inference and training frameworks like ROG that balance performance with resource constraints. Cui also investigates the robustness of deep reinforcement learning agents under adversarial conditions, developing novel attack and defense strategies grounded in policy distribution analysis and adaptive gradient masking. His contributions to visuomotor control and generalist robotic manipulation further demonstrate a broad systems-level vision. With a rapidly growing citation profile across multiple high-impact areas, Cui represents a dynamic voice in next-generation intelligent robotic systems research.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4New Problems in Distributed Inference for DNN Models on Robotic IoT6 citations · 2024
- 5
- 6
- 7
- 8
- 9
- 10