A public project archive for undergraduate research on autonomous manipulation and grasping through imitation learning with visuo-tactile sensing.
Overview
- Research title: Development of Autonomous Manipulation and Grasping Technology based on Imitation Learning with Visuo-Tactile Sensing
- Korean title: 시촉각 센싱기반 모방학습에 의한 자율조작파지 기술개발
- Period: February 2026 – August 2026
- Affiliation: Control and Optimization Laboratory, Hanyang University
- Project type: Undergraduate research
- Role: Expert-data generation and reward review for a KETI Allegro dual-arm cashier policy
- Main tools: Apple Vision Pro, MuJoCo, VisionProTeleop
Photo
Research Context
The work investigated data generation and reward review for an Allegro dual-arm cashier manipulation policy. The project used Apple Vision Pro teleoperation and MuJoCo simulation to support expert demonstration collection and policy-development experiments.
Contributions
- Used the VisionProTeleop project with Apple Vision Pro and MuJoCo for teleoperation-based data generation.
- Generated expert demonstrations for the KETI Allegro dual-arm cashier task.
- Reviewed reward behavior during policy experiments to check whether the reward signal aligned with the intended manipulation and grasping task.
- Recorded this public overview separately; experimental photos and additional data-collection materials have been maintained in a dedicated private archive.
Technical Keywords
Imitation learning, autonomous manipulation, grasping, visuo-tactile sensing, teleoperation, Apple Vision Pro, MuJoCo, Allegro Hand, dual-arm manipulation, expert demonstrations, reward design