EchoView.ai

Project Overview

Collaborators: Nicholas Hardy, Hassan Hijazi, Marybel Boujaoude, Riya Deokar, Jazmyn Walker

Date: 2023-09-01 - 2024-05-01

For my Senior Capstone Project, I chose to put forward a personal project rather than work for an existing company. This was where EchoView.ai was born.

EchoView.ai is dedicated to helping unlock a world of communication, collaboration and human connection. We believe that everyone should hold the key to understanding spoken interactions, so we can foster community, learning, and innovation

Key Features

• Real-time speech-to-text transcription displayed directly in the user's field of view.
• Directional Microphones to focus on the speaker's voice and reduce background noise.
• Bluetooth connectivity to pair with smartphones and other devices.
• Very low prototype cost ($28.60 unit price) to ensure affordability.
• Over 90% accuracy in speech-to-text transcription until 70dB.

Implementation Details

• ESP32-C3 microcontroller chosen for its processing power, wireless capabilities and efficiency. This was used to establish the BLE connection with the iOS application.
• SSD1306 OLED display used to display the real-time transcription into a bird bath optical combining system to acheive required 20cm focal length.
• iOS mobile application developed using Swift to connect to the AR glasses via BLE and process the transcription requests using the on device Siri module.
• Custom 3D printed housing that holds the ESP32-C3, OLED display, mirrors and microphone array. The battery is connected to the ESP-32 via a braided USB-C cable that is routed through the frame and out the back of the glasses.

Supporting Artifacts