Multi-Level-Secure-Authentication-System

S1 POE project

The project is for a better aid for factors that need very effiecient and strong security

📜 Prabhāsa Mantra — A Dual-Channel, Multi-Factor Authentication System

High-Security Hardware + Software Authentication Platform

Prabhāsa Mantra is a custom multi-factor authentication system designed for high-security entry control. It combines:

1. Facial Recognition (Laptop, Python + OpenCV)

2. Cryptographic TOTP Mantra transmitted through:

   • Light channel (color flashes via phone screen)
   • Sound channel (high-frequency tones)

The authentication lock is implemented on an ESP32-S3 microcontroller, which receives the mantra via sensors, verifies its integrity, computes its own TOTP, and grants/denies access.

This repository contains all hardware, software, and protocol details required to fully build, flash, and run the system.

---

🔒 Project Overview

The system uses two layers of authentication:

Factor 1 — Face Authentication (Laptop)

• Python script using OpenCV (LBPH recognizer)
• Runs continuously, scanning the camera feed
• When an authorized face is detected → sends LISTEN_FOR_MANTRA to the ESP32

Factor 2 — Mantra Authentication (ESP32-S3)

After receiving the signal:

1. ESP32 activates:

   • TCS34725 Color Sensor
   • MAX9814 Microphone (FFT-based audio decoding)

2. Waits for the user to transmit the Prabhāsa Mantra from the mobile app:

   • 6-digit TOTP (SHA-256, 30s window)
   • Transmitted first as color flashes, then as sound tones

3. ESP32 decodes both sequences

4. Validates:

   • Light == Sound consistency
   • Light == ESP32’s own internally generated TOTP

If valid → sends ACCESS_GRANTED Otherwise → sends ACCESS_DENIED

---

🧠 Technology Stack

Hardware

• ESP32-S3 DevKitC-1
• TCS34725 RGB Color Sensor
• MAX9814 Microphone + Automatic Gain Control
• USB-Serial connection to laptop

Software

• Flutter App (Android/iOS)

  • Displays TOTP
  • Emits Light+Sound Mantra

• Firebase

  • Authentication (email + password)
  • Firestore user profiles

• ESP32 C++ Code (Arduino Framework)

  • TOTP (HMAC-SHA256 using mbedTLS)
  • FFT audio classification
  • TCS34725 color classification
  • Serial commands + state machine

• Python Script (Laptop)

  • OpenCV LBPH face recognition
  • Serial communication with ESP32

---

🔑 Shared Secret & TOTP Specification

Base32 Secret Key:

GAXG2435ONV6Y3T7MFQZ227VMFQ4W4ZKWSKCA53KPOJ4NEYLTMZVKGIY=

TOTP Parameters

Parameter	Value
Digits	6
Hash	SHA-256
Step	30 seconds
Window	Strict (0 tolerance)

---

📡 Mantra Transmission Protocol

📍 1. Light Channel

6 color flashes, each:

• 500 ms color
• 200 ms black gap

Digit	Color	Hex
0	Red	#FF0000
1	Green	#00FF00
2	Blue	#0000FF
3	Yellow	#FFFF00
4	Magenta	#FF00FF
5	Cyan	#00FFFF
6	Orange	#FFA500
7	Purple	#800080
8	White	#FFFFFF
9	Turquoise	#40E0D0

---

🔊 2. Sound Channel

6 high-frequency tones, each:

• 500 ms tone
• 200 ms silence

Digit	Frequency
0	12000 Hz
1	12500 Hz
2	13000 Hz
3	13500 Hz
4	14000 Hz
5	14500 Hz
6	15000 Hz
7	15500 Hz
8	16000 Hz
9	16500 Hz

---

🧩 Serial Protocol (Laptop ↔ ESP32)

Direction	Message	Meaning
ESP32 → Laptop	TIME_REQUEST	ESP32 asks laptop for current epoch
Laptop → ESP32	TIME_RESPONSE <epoch>	Laptop sends UNIX time (sec)
Laptop → ESP32	LISTEN_FOR_MANTRA	Face recognized; begin mantra capture
ESP32 → Laptop	ACCESS_GRANTED	Lock opens
ESP32 → Laptop	ACCESS_DENIED	Lock stays closed

---

🔌 Hardware Wiring

TCS34725 → ESP32-S3

TCS34725 Pin	ESP32-S3 Pin
VIN	3.3V
GND	GND
LED	3.3V
SDA	GPIO 8
SCL	GPIO 9

MAX9814 → ESP32-S3

MAX9814 Pin	ESP32-S3
VDD	3.3V
GND	GND
OUT	GPIO 4 (ADC1_CH3)

---

🧭 System Flowchart

      +-----------------+
      |    Laptop       |
      |  Face Detect    |
      +--------+--------+
               |
               | LISTEN_FOR_MANTRA
               |
      +--------v--------+
      |    ESP32        |
      |   State: LISTEN |
      +--------+--------+
               |
     -------------------------
     |                       |
 +---v---+             +-----v-----+
 | Light |             |  Sound    |
 | Decode|             |  Decode   |
 +---+---+             +-----+-----+
     |                       |
     -----------+------------
                 |
        +--------v--------+
        |  Compare codes  |
        +--------+--------+
                 |
          +------v------+
          |  Generate    |
          |    TOTP      |
          +------+-------+
                 |
      +----------v-----------+
      |   Access Granted?    |
      +----------+-----------+
                 |
     Yes ------------------- No
     |                       |
 ACCESS_GRANTED      ACCESS_DENIED

---

🛠 Build & Run Instructions

ESP32-S3 (Arduino IDE / PlatformIO)

1. Install Espressif ESP32 Arduino Core

2. Select board: ESP32S3 Dev Module

3. Install libraries:

   • Adafruit_TCS34725
   • ArduinoFFT

4. Wire hardware as shown above

5. Flash main.cpp to the ESP32

---

Laptop (Python, Conda Environment)

Install dependencies:

conda install -c conda-forge opencv numpy
pip install pyserial opencv-contrib-python

Run the script:

python main.py

Place enrollment images:

/enroll
   face1.jpg
   face2.jpg

---

📱 Flutter App

Features:

• Firebase Login

• Biometric Unlock

• TOTP generation using above secret

• “TRANSMIT” button to emit:

  • Light flashes (full-screen color)
  • Audio tones (Ultrasonic band)

---

🧪 Testing Procedure

1. Run Python script → ensure “TIME_REQUEST” appears

2. Bring face into camera → wait for recognition

3. When instructed, transmit mantra from phone:

   • Hold phone screen 1–5 cm from color sensor
   • Hold speaker close to MAX9814

4. Check ACCESS_GRANTED / ACCESS_DENIED on laptop

---

🛡 Security Features

• Two-factor authentication (biometric + cryptographic)
• Offline, air-gapped transmission channels
• TOTP SHA-256 with strict window consistency
• Secret never transmitted over air or serial
• Sound + Light redundancy to prevent spoofing
• Laptop provides authoritative time sync

---

📂 Repository Structure

/
├── firmware/
│   └── main.cpp                # ESP32-S3 Firmware
├── laptop/
│   └── main.py                 # Face recognition + serial control
├── flutter_app/                # Mobile app source (if included)
├── hardware/
│   ├── wiring_diagram.png
│   └── notes.md
└── README.md

---

🤝 Contributing

Contributions are welcome! Feel free to open:

• Issues
• Pull Requests
• Feature suggestions

---

📄 License