HANG ZHAO

Hero

Techsure

Test

2025 / 6
Designer System Engineer Interaction Designer Fabricator
Python C++ Edge AI Embedded Processor PCB Design Haptics
Undergraduate Thesis Project Tsinghua University — Academy of Arts & Design Advisor: Prof. Mi Haipeng
Hang
Overview

Tuchsure is a multisensory interaction system designed for visually impaired users to help them perceive risks they cannot see. By integrating object recognition, edge inference, audio feedback, and tactile vibration, the system provides an intuitive, trust-building experience that reduces anxiety and increases safety and autonomy in daily life.

The device captures environmental images through an embedded camera, processes object-level risk using an RK3588 edge-computing module, and communicates risk levels through differentiated vibration and audio cues.

Img
Problem Space: “Not Seeing Risk is the Real Risk.”

Visually impaired users do not only lose access to visual information—they lose access to risk-awareness. Traditional assistive tools help identify obstacles, but fail to interpret object-level danger (e.g., scissors vs. cup).

Tuchsure aims to restore a sense of security through weak-perception safety design: combining multimodal sensory cues to form reliable, human-centered risk awareness.

Img
Exhibition
Img
局部
Img
Img
System Architecture: Image Recognition + Audio + Tactile Feedback

The system is built around an integrated perception module:

  • ESP32-CAM for image capture
  • RK3588 for YOLOv8 risk inference (mAP50 = 0.984)
  • Customized PCB
  • Vibration motor (finger-mounted)
  • Audio notification module (palm-mounted)
  • Frosted acrylic enclosure & ergonomic grip

Pipeline:

1) Image Capture → 2) Edge Inference → 3) Risk Mapping → 4) Vibration + Audio Feedback

Img
Interaction Design: Natural Interaction from Form

The device is designed around three intuitive actions:

  • Hold (form-follows-finger ergonomics)
  • Feel (localized vibration on fingertip—highest neural density)
  • Listen (fast auditory recognition of object category)

This multimodal feedback aligns with Baddeley’s working memory model, enhancing rapid recall and situational safety.

Feedback Mechanism: Encoded Multisensory Risk Levels

Three vibration modes communicate graded risks:

Audio provides semantic confirmation, while vibration offers spatial, fast-response cues.

Exhibition & Experience Design

The exhibition includes:

  • Main installation: demonstration of the wearable device
  • Three experimental devices: tactile coding test stations
  • Visual panels & demonstration video
  • Interactive flow guiding visitors through recognition → vibration → interpretation

This creates a coherent narrative around risk perception and multisensory interaction.

Guiding Question

How can multimodal interaction empower visually impaired users to perceive hidden risks in daily life?

Img
Img
Img
Img
Test section

Hi this is a test, Tuchsure is a multisensory interaction system designed for visually impaired users to help them perceive risks they cannot see. By integrating object recognition, edge inference, audio feedback, and tactile vibration, the system provides an intuitive, trust-building experience that reduces anxiety and increases safety and autonomy in daily life.

Column Image