Chenxu Li

News

• May 2026

  Our paper RadioShock: Over-the-Air Adversarial Attacks on Wireless Communication was accepted by IEEE Transactions on Dependable and Secure Computing (TDSC).

Research

Asterisk (*) denotes equal contribution; dagger (†) denotes the corresponding author.

MoMaStage: Skill-State Graph Guided Planning and Closed-Loop Execution for Long-Horizon Indoor Mobile Manipulation

Chenxu Li*, Zixuan Chen*, Yetao Li, Jiapeng Xu, Hongyu Ding, Jieqi Shi, Jing Huo, Yang Gao

arXiv preprint arXiv:2603.08383, 2026.

We propose MoMaStage, a structured vision-language framework for long-horizon indoor mobile manipulation that grounds VLM-based planning within a topology-aware Skill-State Graph and Hierarchical Skill Library, enabling logically consistent, closed-loop skill execution and semantic replanning without explicit scene mapping, thereby substantially improving robustness, planning efficiency, and task success across simulation and real-world environments.

[arXiv] [Project] [Code]

V-Dreamer: Automating Robotic Simulation and Trajectory Synthesis via Video Generation Priors

Songjia He*, Zixuan Chen*, Hongyu Ding, Dian Shao, Jieqi Shi, Chenxu Li, Jing Huo, Yang Gao

arXiv preprint arXiv:2603.18811, 2026.

We present V-Dreamer, a fully automated framework that converts natural language instructions into open-vocabulary, simulation-ready manipulation environments and executable robot trajectories by combining physically grounded 3D scene generation with video-driven motion priors and visual-kinematic alignment, enabling scalable data synthesis and robust sim-to-real generalization for robotic manipulation.

[arXiv]

RadioShock: Over-the-Air Adversarial Attacks on Wireless Communication

Wenhao Li, Chenxu Li, Guoming Zhang, Zhijian Huang, Gang Qu, Xiuzhen Cheng, Jun Luo, Pengfei Hu†

IEEE Transactions on Dependable and Secure Computing (TDSC), 2026.

We present RadioShock, a practical over-the-air adversarial attack framework for wireless communication systems that leverages dynamic channel-state adaptation and universal compact perturbation generation to achieve covert real-world attacks, significantly degrading the performance of deep learning–based communication models under realistic channel conditions.

[IEEE Xplore]