Skip to main content

5. Integration

Smart Glasses Kit – System Integration Documentation

Overview​

The Smart Glasses Kit is a modular, attachable device that can be mounted onto conventional eyeglasses to bring them smart capabilities. This kit supports voice interaction and visual understanding, enabling use cases such as:

  • “What’s the weather today?” – The glasses respond via voice.
  • “What’s in front of me?” – The glasses analyze and report what they see.

This document outlines the full system integration, including hardware, communication structure, software pipeline, and the bill of materials (BOM).

System Architecture​

Hardware Components​

1. Server​

  • A local PC acts as the processing server, running TTS and STT models.
  • A LLM model (e.g., OpenAI Whisper, OpenAI GPT) is used for conversational AI and image analysis.

2. Glasses Kit Modules​

ComponentUnit Price (RMB)Unit Price (USD, approx)
XIAO ESP32S3 with Sense (cam + mic)—$14.99
SmartGlassesKit PCB (5 for ¥20)¥4~$0.56
LED (5 for ¥5)¥1~$0.14
Button (5 for ¥5)¥1~$0.14
GC9107 0.85" LCD¥29~$4.04
MAX98357 Audio Amplifier¥10~$1.39
Small Speaker¥5~$0.70
Beam Splitter Prism¥20~$2.78
Lion Battery¥10~$1.39
Custom TPU 3D-Printed Case—~$1.50 (estimated)

Total Estimated BOM Cost: ~$27.24

[Insert Illustration Here – "Exploded View of Smart Glasses Kit"]​

(Use the exploded-view diagram with labeled components: case, prism, speaker, MAX98357, LCD, ESP32S3, and PCB)

Hardware Integration​

The components are stacked in a compact form:

  1. Core Controller: XIAO ESP32S3 receives audio and video input, and transmits data to server.
  2. Display: GC9107 shows simple visual results.
  3. Audio Output: MAX98357 + speaker play voice responses.
  4. Interaction: Button toggles modes or confirms actions.
  5. Prism: Reflects display output into user's field of view.
  6. 3D-Printed Case: Provides enclosure and mounting.

Software Workflow​

Use Cases​

  1. Weather Inquiry

    • User speaks: "What’s the weather?"
    • ESP32S3 records → sends to server → server fetches weather → responds via TTS

  2. Visual Scene Detection

    • User says: "What’s in front of me?"
    • ESP32S3 captures photo → server runs CV model → responds with TTS + result image

Communication Protocol​

  • Wi-Fi MQTT or WebSocket connection to the server
  • STT and CV processed on server side
  • Results sent back as structured JSON:
{
  "imgage": "base64",
  "text": "There are three people ahead of you."
}

Assembly Guide​

  1. Solder components on the SmartGlassesKit PCB.
  2. Mount XIAO ESP32S3, LCD, and speaker.
  • Connect Sense to XIAO ESP32S3
  • COnnect Speaker to MAX98357
  • Connect LCD to SmartGlassesKit PCB(designed by me)
  • Connect MAX98357 and XIAO ESP32S3 to SmartGlassesKit PCB.
  • Hero shot of the assembled kit.
  1. Connect beam splitter prism in line with LCD.
  2. Enclose everything in 3D-printed TPU case.
  3. Attach kit to glasses frame with clips.

Testing​

Resources​