Embodied AI and planning
Organize geometric and semantic evidence in assembly knowledge graphs, use LLMs to propose actions, and verify and refine them through physical simulation.
- Assembly Knowledge Graph
- LLM-guided Planning
- Physics Verification
ZITENG WANG · WANG ZITENG
Ziteng Wang 王子腾
PhD student in Computer Science · Beihang University
I study how people perceive, interact and act in virtual and physical environments. My current work spans embodied AI, VR/AR, human–computer interaction, and knowledge-driven assembly and disassembly planning.
06
Published papers
VR · HCI
Immersive interaction
Embodied AI
Current research focus
RESEARCH
Connecting perception, reasoning and action through interaction.
My work began with VR input and multi-object manipulation and now extends to embodied data, knowledge representation and executable planning. I develop methods and systems, run user studies and evaluate physical feasibility.
VR/AR and human–computer interaction
Design touch, gaze and bare-hand interactions for text entry, multi-object selection, spatial arrangement and collaboration.
- Spatial Interaction
- Eye & Hand Input
- User Study
Perception and systems
Build multi-camera capture systems, 3D hand perception, robot retargeting, on-device Android applications and AI agent platforms.
- 3D Hand Perception
- Robot Retargeting
- Full-stack Prototyping
CURRENT FOCUS
Knowledge-driven disassembly planning
How can we make language-model disassembly plans physically executable?
I extract contact relationships and geometric attributes from CAD assemblies and enrich them with part functions and assembly semantics. A planner proposes sequences from structured states; a simulator checks collisions and blocking, then returns failure evidence for the next reasoning step.
CAD Assembly
Knowledge Graph
LLM Planning
Physics Check
PUBLICATIONS
Publications
Six published papers on immersive interaction and multimodal perception. Each page explains the method, experimental evidence and citation details.
Collaborative Disassembly for Multiple Users in Virtual Reality
Dynamic Parallel Disassembly and Tasking connects simulation-validated planning, parallel task sets, and proximity-aware allocation for collaborative VR.
DSPT: Disassembly Sequence Planning Transformer for Interaction Guidance in VR
A Transformer uses disassembly-sequence and part-history features to predict removal candidates for interactive VR guidance.
MOA: Efficient Scene-Aware Multi-Object Arrangement in VR
Scene-aware selection and auxiliary manipulation structures support controller-free arrangement of many occluded objects in VR.
Align-Box: Enable Fast Bare-Hand Multi-Object Alignment in VR
A two-box interaction metaphor maps an object enclosing box to a proxy box for bare-hand group-based and object-based alignment in VR.
Multimodal Contrastive Learning for Cybersickness Recognition Using Brain Connectivity Graph Representation
Brain connectivity graph representations align EEG, video, and motion for multimodal cybersickness recognition.
FanPad: A Fan Layout Touchpad Keyboard for Text Entry in VR
A curved, customizable QWERTY layout follows thumb movement on two VR controller touchpads.
SELECTED SYSTEMS
Selected systems and demos
ANDROID · COMPUTER VISION
Android Sports Assistant
A mobile prototype that connects motion input, visual recognition and immediate feedback for sports activities.
HAND TRACKING · DEVICE TEST
Pico 4 Pro Hand Tracking
Test the headset's native hand tracking during interaction, examining its tracking range, occlusion behavior and stability.
OPEN SOURCE
More open-source projects
BACKGROUND
Education and experience
Alongside research, I build systems that put methods into practice.
2025.09 — present
EDUCATION
Beihang University · Computer Science and Technology
PhD student at the School of Computer Science and Engineering, State Key Laboratory of Virtual Reality Technology and Systems. Research in embodied AI, VR/AR and HCI.
2021.09 — 2025.06
EDUCATION
Beihang University · Virtual Reality Technology
Bachelor's degree, GPA 3.78 / 4.00. Outstanding Graduate of Beihang University.
2024 — present
ENGINEERING
AI glasses and embodied data platforms
Engineering work on Android AI glasses, multi-sensor capture, data quality checks, camera calibration, hand perception and robot retargeting.
2023 — 2025
TEACHING
Teaching assistant
Teaching assistant for Data Structures and Artificial Intelligence.
ARCHIVE
Notes and experiments
Course notes, experiment logs and essays, preserved in their original languages.