Yihan Li

+86 18615276438 | liyihan.xyz@gmail.com

Personal Website: liyihan.xyz

|| Education

Sun Yat-sen University

English Major, School of Foreign Language Sep. 2023 – Jun. 2027 (Expected)

Overall GPA 3.8/5.0

Computer Science and Technology Minor, School of Computer Science and Engineering

Mathematics: Linear Algebra, Calculus, Discrete Mathematics, Mathematical Modeling.

Computer Science: Data Structures & Algorithms, Natural Language Processing (NLP), Programming Design.

|| Publications

RAPID Hand: Robust, Affordable, Perception-Integrated, Dexterous Manipulation Platform for Embodied Intelligence

Zhaoliang Wan, Zetong Bi, Zida Zhou, Hao Ren, Yiming Zeng, Yihan Li, Cao He, Lu Qi, Xu Yang, Ming-Hsuan Yang, Hui Cheng

NeurIPS 2025 Conference

Unordered Landmark Visual Navigation

Hao Ren, Junzhe Zhu, Yihan Li, Zetong Bi, Le Zheng, Zhi Li, Yiqing Yuan, Zhaoliang Wan, Lu Qi, Hui Cheng

CVPR 2026 Conference Submission

|| Research Experience

Research on Perception-Integrated Dexterous Manipulation (RAPID Hand)

Research Assistant, Prof. Hui Cheng, Nov. 2024 – Jun. 2025

• Developed real-time low-level drivers for a custom 20-DoF anthropomorphic hand, achieving a 1kHz control loop and ensuring millisecond-level synchronization for multi-modal data.

• Designed an automatic calibration algorithm to align visual-tactile sensory data, stabilizing perception latency within 7ms and providing high-quality datasets for the lab’s NeurIPS submission.

• Constructed a high-fidelity MuJoCo simulation environment for Sim2Real verification, supporting the training of Diffusion Policies that achieved SOTA performance in complex grasping tasks.

TopoVisNav: Visual Topological Mapping with Dynamic Pruning

Core Developer & Co-author, Prof. Hui Cheng, Nov. 2024 – Present

• Designed TopoVisNav, a coarse-to-fine visual mapping framework integrating global retrieval (MegaLoc) and local geometric verification (LightGlue) for robust place recognition.

• Proposed a dynamic parameter analysis method using K-Means clustering to automatically determine the optimal matching threshold τ, replacing manual tuning.

• Implemented a graph pruning strategy based on Maximum Spanning Forest (MSF) to remove redundant edges while retaining topological connectivity, verified by PRF1 scores against Ground Truth.

LLM-based Automated Annotation for Academic Discourse

(Guangdong Provincial Philosophy and Social Science Planning Project 2025)

Research Assistant, Assoc. Prof. Man Guo, Jan. 2025 – Oct. 2025

• Engineered an automated annotation system using Large Language Models (LLMs), employing advanced Prompt Engineering to extract linguistic projection relations from unstructured text.

• Achieved 85% annotation accuracy through iterative fine-tuning, successfully converting complex academic discourse into hierarchical structured data for large-scale quantitative analysis.

|| Work Experience

Sun Yat-sen University RoboTech Association

Nov. 2023 – Present

President, Competition Team Captain

• Orchestrated the daily operations and training programs of the university’s largest robotics association, managing a multidisciplinary team of over 160+ members.

• Organized the university-level “Smart Space” Unmanned Vehicle Technology Challenge, coordinating logistics, rule-setting, and technical support for 47 participating teams.

• Spearheaded the preparation for major national competitions (RoboMaster, Diansai), overseeing project timelines, budget allocation, and cross-team collaboration, resulting in multiple national awards.

Yun Drone

Jun. 2025 – Oct. 2025

R&D Intern

• Executed the hardware selection and system integration for scientific research UAVs, finalizing the configuration of flight controllers, ESCs, and sensor suites for optimal payload capacity.

• Conducted rigorous flight tests to evaluate stability and maneuverability, utilizing flight logs to identify control anomalies and providing data-driven recommendations for PID tuning.

• Collaborated with the mechanical team to optimize the drone’s structural layout, ensuring electromagnetic compatibility (EMC) and vibration isolation for sensitive sensors.

|| Project Experience

RoboMaster Visual System Development (RMUL)

Visual Group Leader

Mar. 2025 – Jul. 2025

• Engineered a high-performance auto-aiming system on NVIDIA Jetson and Intel NUC, utilizing Extended Kalman Filter (EKF) to predict the trajectory of high-speed erratic targets, achieving a 90%+ hit rate.

• Deployed a YOLO-based armor classification model optimized with TensorRT and OpenVINO, reducing inference latency to 10ms to ensure real-time closed-loop control.

• Implemented a robust communication protocol between the Vision PC and STM32 MCU, ensuring precise gimbal synchronization under high-latency wireless conditions.

Autonomous Navigation and Precise Robotic Arm Pick-and-Place System for Off-Road Vehicles in Complex Scenarios

Project Leader

Mar. 2025 – Jul. 2025

• Focused on 4WD differential and Ackermann steering chassis. Optimized FAST_LIO2 to solve mapping and loop closure issues in degenerate environments; used FAST_LIO2’s high-frequency odometry combined with ICP for efficient and accurate relocalization.

• Optimized the local planner based on TEB to achieve high-speed, precise navigation. Equipped with a 6-DoF robotic arm, utilizing binocular vision point clouds, YOLOv11 algorithm, and PnP solving to achieve precise pick-and-place operations.

UAV Swarm Flight Based on Mesh Networking Communication and GNSS RTK

Core Member

Jul. 2025 – Oct. 2025

• Debugged and deployed the Mesh self-organizing network communication module, establishing a real-time data transmission link between UAVs.

• Collected and validated GNSS RTK positioning data, completing high-precision UAV position calibration, and analyzing positioning error sources by comparing static/dynamic flight RTK results with base station data.

• Co-designed the UAV swarm flight test plan, assisting in multi-UAV coordination tasks (e.g. swarm takeoff/landing, trajectory tracking, dynamic formation changes).

GenreShift: An Intelligent Genre Conversion Expert System Based on Quantitative Linguistic and Large Language Models

Project Leader

Oct. 2024 – Nov. 2025

• Constructed quasi-parallel corpus, data crawling and cleaning, stylometric feature extraction, LLM LoRa fine-tuning, and Mixture of Experts (MoE) system construction.

• Directed the overall project planning; an invention patent has been applied for (Title: “一种基于计量语言学与 LLM 的智能体裁转换方法及系统,” Application No.: 2025105553019).

|| Honors and Awards

National First Prize (Team Leader)

China Robot and Artificial Intelligence Competition

Champion (Team Leader)

National Intelligent Unmanned Systems Application Challenge

National Second Prize (Team Leader)

National Undergraduate Electronics Design Contest

National Second Prize (Visual Group Leader)

RoboMaster University League

Honorable Mention (Team Leader)

Mathematical Contest in Modeling (MCM)

2025

2025

2025

2025

2024

|| Professional Skills

Programming: ROS1/2, Python, C/C++, MATLAB, STM32/Arduino Embedded Development

Skills: Fusion 360/SolidWorks Modeling, EDA Design

Licenses: AOPA Multi-rotor VLOS Pilot, C1 Driver’s License

Languages: Mandarin (Native), English (TEM-4), French

Interests: Badminton, Soccer, Cycling, FPV Drone