Human-machine Interaction Lab
Beihang University
This project addresses the difficulty of describing, comparing and standardizing leather tactile qualities. It develops a hardware system and recognition algorithms for multidimensional leather haptic property measurement, turning experience-based tactile judgment into measurable, repeatable and modelable data.
Project Goals
The project combines Beihang's expertise in haptics, human-machine interaction, virtual reality and robotics with Xingye Leather's industrial resources in materials, manufacturing and market scenarios. The goal is to build an innovation platform for research, development, talent training and technology transfer.
Six Core Directions
Technical Route
The system integrates BioTac and Paxini tactile sensors, industrial cameras, a 3D laser profilometer, a motion platform and robotic fingers. It synchronously captures tactile, visual, surface-profile and contact-state data.
The new architecture uses a Raspberry Pi application layer, an STM32 control layer and peripheral drivers to create a stable acquisition and control chain. Contact position, indentation depth, contact angle and force become measurable, controllable and repeatable variables.
Progress
The first prototype has validated the feasibility of leather haptic measurement. Initial tests on small sample batches showed that several tactile dimensions can be stably extracted while visual and shape data are collected.
The current phase focuses on system reconstruction, multimodal sensing upgrades and robotic-finger integration. The robotic finger can imitate key human touch behaviors, including repeatable contact positioning, posture adjustment and controllable normal force.
Data and Algorithms
The project is moving from single-point tactile measurement toward multimodal fusion of vision, touch and surface geometry. The next stage will build baseline recognition models for compliance, texture, thermal properties and friction, while establishing sample organization and data standards.
Long term, the project will also model tactile preference: not only what a leather feels like, but which tactile features users prefer and why.
Expected Outputs
- A stable leather haptic property measurement platform
- A multimodal data workflow covering touch, vision, surface geometry and contact states
- Quantitative methods for compliance, texture, thermal properties and friction
- A leather tactile digital model library for design and production
- Programmable haptic presentation and virtual leather technologies for VR/AR/MR experiences